Showing 1–50 of 596 results for author: Cao, Q

TileMix: Tile-Centric Mixed-Precision Attention for LLM Inference Acceleration

Authors: Hanzhi Zhang, Qiao Zhang, Qinglei Cao, Heng Fan, Yan Huang, Kewei Sha, Yunhe Feng

Abstract: Long-context prefill in large language models (LLMs) incurs substantial computation and memory traffic because dense self-attention computes quadratic query-key scores. Existing methods either use a uniform low-precision path or select token interactions, leaving spatial precision routing over hardware-aligned score tiles outside fused dense attention. We introduce TileMix, a tile-centric precisio… ▽ More Long-context prefill in large language models (LLMs) incurs substantial computation and memory traffic because dense self-attention computes quadratic query-key scores. Existing methods either use a uniform low-precision path or select token interactions, leaving spatial precision routing over hardware-aligned score tiles outside fused dense attention. We introduce TileMix, a tile-centric precision-routing kernel that makes numerical precision an executable spatial decision over score-tile groups within fused dense attention. TileMix partitions the attention matrix into hardware-aligned score tiles, packs routing decisions into compact bitmasks, and dispatches each tile group through FP16 or INT8 score computation while both paths update a shared online-softmax state. Scalable precision grouping lets each routing bit govern multiple adjacent key tiles, preserving hardware-aligned compute tiles and compact metadata at long contexts. By routing all legal tile groups, TileMix preserves dense token connectivity, requires no training, and supports grouped-query attention, variable-length batches, and INT8 key/value caches. Across LongEval, LV-Eval, and A100 prefill benchmarks on LLaMA, Qwen, and Vicuna, TileMix recovers long-context quality lost under uniform INT8 and improves prefill throughput over FP16, yielding a controllable accuracy-efficiency frontier across model families. The implementation is available at https://github.com/HanzhiZhang-Ulrica/TileMix. △ Less

Submitted 17 August, 2026; originally announced August 2026.

scDNM-VAE enables directly inspectable deep clustering of single-cell RNA-seq data through signed dendritic gating

Authors: Melih Agraz, Deniz Karapinar, Aysel Topsir, Qianying Cao, Erol Egrioglu, Gaurav Choudhary

Abstract: Deep clustering models for single-cell RNA sequencing often assign cells through latent or centroid-based mechanisms that are difficult to inspect. We introduce scDNM-VAE (single-cell Dendritic Neuron Model Variational Autoencoder), a deep clustering framework that combines a variational autoencoder with a dendritic neuron-inspired head. Cluster assignments are governed by learnable signed synapti… ▽ More Deep clustering models for single-cell RNA sequencing often assign cells through latent or centroid-based mechanisms that are difficult to inspect. We introduce scDNM-VAE (single-cell Dendritic Neuron Model Variational Autoencoder), a deep clustering framework that combines a variational autoencoder with a dendritic neuron-inspired head. Cluster assignments are governed by learnable signed synaptic weights and thresholds: the weight sign determines the direction of a gate's response to a latent coordinate, its magnitude controls steepness, and the weight-threshold pair determines the transition location. The trained clustering function can therefore be inspected directly without fitting a post-hoc explanation model. We benchmark scDNM-VAE on four datasets spanning immune, cortical, cardiac, and hematopoietic cells against scVI followed by KMeans and an MLP-DEC ablation. scDNM-VAE performs better than scVI on PBMC3k, comparably on the Human Heart Cell Atlas and Paul15, and worse on Zeisel, while producing biologically coherent marker-gene signatures. Ablating each cluster's three highest-magnitude synaptic dimensions causes numerically greater reassignment than random-dimension ablation across all datasets, but the margins are modest and negligible on Zeisel. These results show that signed dendritic gating supports competitive clustering with a parameter-inspectable decision function, while indicating that decision-relevant information is distributed across the latent space. △ Less

Submitted 17 August, 2026; originally announced August 2026.

Comments: 39 pages, 7 figures, 4 tables. Supplementary information available from the authors

A Unified DINOv2-Based Framework for LVEF Estimation, GLS Dysfunction Classification, and Early Cardiotoxicity Prediction

Authors: Xiaotong Zhang, Mingyue Cui, Qing Cao, Jingming Xia

Abstract: Left ventricular ejection fraction (LVEF) estimation (Task 1), global longitu-dinal strain (GLS)-based dysfunction classification (Task 2), and early cardi-otoxicity prediction (Task 3) provide complementary information for cardio-oncology assessment. LVEF reflects macroscopic ventricular volume chang-es as the clinical standard, whereas GLS captures subtle myocardial defor-mation, indicating subc… ▽ More Left ventricular ejection fraction (LVEF) estimation (Task 1), global longitu-dinal strain (GLS)-based dysfunction classification (Task 2), and early cardi-otoxicity prediction (Task 3) provide complementary information for cardio-oncology assessment. LVEF reflects macroscopic ventricular volume chang-es as the clinical standard, whereas GLS captures subtle myocardial defor-mation, indicating subclinical cardiotoxicity before overt LVEF decline. Fur-thermore, predicting cardiotoxicity from baseline echocardiography prior to treatment enables preventive interventions at an early stage. To address these three tasks, we employ a DINOv2-based framework with task-specific adap-tation and prediction heads. Built upon a frozen foundation encoder, the framework incorporates parameter-efficient Low-Rank Adaptation (LoRA) and temporal aggregation to learn task-specialized representations, ensuring robust generalization. Crucially, during inference, it operates in a fully cycle-detection-free and phase-free manner, requiring neither cardiac cycle seg-mentation nor explicit End-Diastolic/End-Systolic (ED/ES) annotations. Ad-ditionally, we introduce an ED/ES-guided 2D/3D hybrid multi-view regres-sion model specifically to optimize Task 1. On a patient-level split containing 1,203 training videos from 237 patients and 300 validation videos from 59 independent patients, the DINOv2-based framework achieved a mean abso-lute error (MAE) of 5.03% for Task 1, an AUC-ROC of 76.48% for Task 2, and an AUC-ROC of 70.26% for Task 3. For Task 1, the specialized ED/ES-guided model further improves performance, achieving an MAE of 4.64%. This framework demonstrates the effectiveness of foundation model repre-sentations across diverse cardio-oncology tasks and the additional benefit of physiology-guided modeling for accurate LVEF estimation. △ Less

Submitted 13 August, 2026; originally announced August 2026.

Comments: Accepted as an oral at the EchoRisk Challenge Workshop, MICCAI 2026

StorySpark: Module-wise Evolutionary Search for Story Premise Generation

Authors: Yang Yang, Zining Zhong, Qian Cao, Jindong Li, Boyun Xu, Kaishen Yuan, Menglin Yang, Yutao Yue

Abstract: A story premise is the creative spark from which a full narrative can grow. Yet LLM-based story generation has mostly emphasized later-stage planning, controllability, coherence, and prose expansion, while premise-level ideation remains comparatively underexplored. We introduce StorySpark, a module-wise evolutionary search framework for story premise generation. StorySpark operates over interpreta… ▽ More A story premise is the creative spark from which a full narrative can grow. Yet LLM-based story generation has mostly emphasized later-stage planning, controllability, coherence, and prose expansion, while premise-level ideation remains comparatively underexplored. We introduce StorySpark, a module-wise evolutionary search framework for story premise generation. StorySpark operates over interpretable narrative modules such as background, persona, event, ending, and twist, treating each active module not as a static field to fill once, but as a local search space conditioned on the partial premise built so far. For each module, it generates alternatives, evaluates them in context, refines them through feedback-driven mutation and recombination, preserves complementary strengths with Pareto-guided selection, and reallocates frontier capacity to balance branch coverage with promising directions. Multi-view automatic and human evaluations show that StorySpark produces stronger final premises than competitive baselines, with especially consistent gains in originality; when expanded with the same story writer, its premises also lead to higher-quality downstream stories while maintaining completeness, fascination, and diverse usable narrative directions. △ Less

Submitted 2 June, 2026; originally announced August 2026.

Comments: 26 pages, 7 figures

Hybrid-Policy Self-Editing for Composable Unstructured Knowledge Editing

Authors: Tianci Liu, Zihan Dong, Tianchun Li, Yi-Chung Chen, Qiming Cao, Xingchen Wang, Shiyang Wang, Zichen Miao, Linjun Zhang, Haoyu Wang, Jing Gao

Abstract: Large language models (LLMs) achieve remarkable performance across natural language tasks, yet they are trained on static corpora and their knowledge quickly becomes outdated in a fast-changing world. This motivates knowledge editing (KE), which updates specific knowledge in an LLM without changing unrelated others. Recent works move from structured knowledge triples toward unstructured KE (UKE),… ▽ More Large language models (LLMs) achieve remarkable performance across natural language tasks, yet they are trained on static corpora and their knowledge quickly becomes outdated in a fast-changing world. This motivates knowledge editing (KE), which updates specific knowledge in an LLM without changing unrelated others. Recent works move from structured knowledge triples toward unstructured KE (UKE), where the edit is a free-form passage that may state multiple facts at once. Nonetheless, existing editors inject such a passage yet fail to use it: the edited model can recall the passage, but can neither answer atomic questions about its facts nor compose them into multi-hop reasoning. We attribute this missing property, which we term composability, to editors' passive reliance on the fixed passage as the sole learning source. In response, we cast editing as a proactive self-distillation from a privileged in-context state of the same model, which requires no external supervision. We further reveal that due to the novelty of the injected knowledge, the pre-edited model's own rollouts rarely cover it, which limits the effectiveness of pure on-policy distillation. To close this gap, we propose HPSE, which builds a hybrid rollout that steps in to place missing facts onto the student's own trajectory precisely where its coverage fails, while staying on-policy elsewhere. We theoretically analyze HPSE's advantage over pure on-policy distillation, and empirically establish its plug-and-play improvements across four LLM backbones and two KE editors under various scenarios. △ Less

Submitted 12 August, 2026; originally announced August 2026.

SearchAuditor: Auditing and Attributing Failures in Long-Horizon Search Agents

Authors: Zhixiang Liang, Yifei Liu, Yidan Huang, Haozhe Zhao, Beichen Huang, Jiaqi Wang, Nan Duan, Qiong Cao

Abstract: Deep search agents tackle challenging questions through long-horizon web interactions, a process that is both complex and fragile: small reasoning errors may propagate through long, noisy trajectories into fluent but incorrect answers. Diagnosing such failures is difficult, requiring the manual inspection of extremely long execution traces, which could be beyond human capacity. We therefore introd… ▽ More Deep search agents tackle challenging questions through long-horizon web interactions, a process that is both complex and fragile: small reasoning errors may propagate through long, noisy trajectories into fluent but incorrect answers. Diagnosing such failures is difficult, requiring the manual inspection of extremely long execution traces, which could be beyond human capacity. We therefore introduce SearchAuditBench, a benchmark that evaluates whether LLM auditors can localize, attribute, and repair these failures, thereby reducing the human burden. SearchAuditBench comprises 1,243 failed trajectories, averaging 73.1 messages and 65.1K tokens, collected from eight open-weight models on five deep-search benchmarks, each expert-annotated with the critical error step, a search-specific root cause, and a reference repair with grading rubrics. We further propose SearchAuditor, a multi-perspective auditing framework that effectively localizes, attributes, and repairs search-agent failures through evidence-grounded adjudication. Experimental results show that even the strongest baseline, when powered by a frontier model like GPT-5.5, attains only a 26.6% end-to-end pass rate. In contrast, our SearchAuditor consistently outperforms all baselines across different frontier models, achieving an end-to-end pass rate of 32.3%, and resuming failed runs with its repairs enables agents to better recover from errors. △ Less

Submitted 5 August, 2026; originally announced August 2026.

SearchMaster: Grounded and Regulated Self-Play for Search Agents

Authors: Wentao Tan, Qiong Cao, Jiaqi Wang, Nan Duan

Abstract: Training LLM-based search agents requires high-quality search data: tasks that demand genuine multi-hop retrieval and trajectories that use search tools effectively. Existing pipelines often depend on human-written tasks, expert demonstrations, or stronger teacher models. We present SearchMaster, a self-play framework that trains a single LLM from search tasks it generates, solves, and verifies in… ▽ More Training LLM-based search agents requires high-quality search data: tasks that demand genuine multi-hop retrieval and trajectories that use search tools effectively. Existing pipelines often depend on human-written tasks, expert demonstrations, or stronger teacher models. We present SearchMaster, a self-play framework that trains a single LLM from search tasks it generates, solves, and verifies in a local search environment. The key challenge is that self-generated tasks and rollouts can yield misleading signals: pseudo multi-hop questions, success-rate difficulty estimates that ignore search depth, and rollouts with excessive opening but little targeted evidence acquisition. SearchMaster addresses these failure modes with three controls. An Evidence-Chain Generator (ECG) grounds task generation in explicit cross-document evidence chains to reduce pseudo multi-hop questions. A Search-Depth Reward (SDR) scores task difficulty by the search depth of successful rollouts rather than success rate alone, keeping retained tasks search-intensive. An Over-Opening Penalty (OOP) regulates tool use by discouraging excessive document opening, avoiding long but shallow browsing. Verified Proposer and Solver rollouts are then jointly optimized with GRPO. Across six deep-search benchmarks, SearchMaster improves a Qwen3.5-9B backbone from 38.19% to 51.52% average accuracy, with a 30.1-point gain on BrowseComp-Plus. These results show that grounded and regulated self-play can provide effective search-agent training data without human-labeled QA pairs or expert demonstrations. The code is available at https://github.com/WentaoTan/SearchMaster. △ Less

Submitted 3 August, 2026; originally announced August 2026.

NLSM amplitudes from a quartic two-derivative theory

Authors: Qu Cao, Zhenqi Han, Fan Zhu

Abstract: We revisit the well-known nonlinear sigma model (NLSM), an effective field theory describing the scattering of $\mathrm{SU}(N)$ Goldstone bosons and characterized by an infinite tower of two-derivative interactions. We introduce a local scalar Lagrangian involving two scalar fields $ψ^\pm$ carrying opposite "polarities", whose interacting part consists of a single polynomial quartic two-derivative… ▽ More We revisit the well-known nonlinear sigma model (NLSM), an effective field theory describing the scattering of $\mathrm{SU}(N)$ Goldstone bosons and characterized by an infinite tower of two-derivative interactions. We introduce a local scalar Lagrangian involving two scalar fields $ψ^\pm$ carrying opposite "polarities", whose interacting part consists of a single polynomial quartic two-derivative operator, and prove that it reproduces planar NLSM amplitudes at all loop orders. This quartic two-derivative formulation reveals a previously hidden simplicity of the NLSM: its Feynman rules involve only a single quartic interaction vertex, making the Adler zero and the leading double-soft factor manifest at all loop orders on generalized cuts and significantly improving the efficiency of high-multiplicity computations. We also suggest a possible analogous description of the NLSM +$ φ^3 $ theory in terms of a finite tower of local interactions. △ Less

Submitted 29 July, 2026; originally announced July 2026.

Comments: 12 pages, 14 figures

mmSimPrior: Learning Simulation Priors for Data-Efficient and Generalizable Real-World Radar-based Human Motion Reconstruction

Authors: Cheng Guo, Qiming Cao, Shengkai Xu, Haoyu Xie, Kaixiang Su, Pu Wang, Hongfei Xue

Abstract: Millimeter-wave (mmWave) radar enables privacy-preserving and illumination-robust human motion reconstruction, but training generalizable models typically requires costly paired radar-motion recordings. Simulation can scale such supervision, yet even physics-based simulators cannot fully reproduce real-world multipath, clutter, hardware-specific response statistics, or distance-dependent resolutio… ▽ More Millimeter-wave (mmWave) radar enables privacy-preserving and illumination-robust human motion reconstruction, but training generalizable models typically requires costly paired radar-motion recordings. Simulation can scale such supervision, yet even physics-based simulators cannot fully reproduce real-world multipath, clutter, hardware-specific response statistics, or distance-dependent resolution degradation, leaving a sim-to-real gap. We present mmSimPrior, a simulation-pretrained framework that factorizes transferable knowledge into signal, motion, and radar-to-motion mapping priors. To learn transferable signal and motion priors, we pretrain a multimodal radar encoder with a physics-informed domain-randomization curriculum designed to mitigate the sim-to-real gap by approximating real-world propagation- and acquisition-level variations, while a joint-temporal tokenizer learns a discrete prior over plausible human motion. A dual-mode mapping module predicts either motion-code distributions for structurally constrained zero-shot reconstruction or continuous motion parameters for flexible adaptation from limited real data. We further construct a 4.2M-frame, 31K-sequence dataset suite and introduce a No-Overlap Setting that prevents any exact subject-environment-location-motion tuple from appearing in both the adaptation and test sets. Experiments on mmSimPrior-Real and RT-Pose demonstrate consistent gains: with only 24 paired real sequences, mmSimPrior-Reg reduces MPJPE by 24.7-39.0% over the strongest baseline across the three environments, while mmSimPrior-Cls reduces zero-shot MPJPE by 8.5% without fine-tuning. △ Less

Submitted 1 August, 2026; v1 submitted 24 July, 2026; originally announced July 2026.

Comments: 19 pages, 11 figures, including supplementary material. Project page: https://ch3ngguo.github.io/mmsimprior/

The Second LoViF 2026 Challenge on Real-World All-in-One Image Restoration: Methods and Results

Authors: Xiang Chen, Hao Li, Jiangxin Dong, Jinshan Pan, Xin Li, Hongbo Ding, Junpeng Jiang, Xingyu Qiu, Yilian Zhong, Yuxiang Chen, Shibo Yin, Zixuan Huang, Yushun Fang, Xilei Zhu, Yahui Wang, Chen Lu, Xiaodong Zhou, Qingyue Cao, Changwei Gong, Jingyun Liu, Xingchen Yi, Hansen Shi, Ruiyi Liu, Jirui Xie, Tao Liu , et al. (67 additional authors not shown)

Abstract: This paper presents a review of the second LoViF Challenge on Real-World All-in-One Image Restoration. The challenge aims to advance unified image restoration under diverse real-world degradation conditions, including blur, low-light, haze, rain, and snow. It provides a common benchmark for evaluating the restoration accuracy, robustness, and generalization capability of models across multiple deg… ▽ More This paper presents a review of the second LoViF Challenge on Real-World All-in-One Image Restoration. The challenge aims to advance unified image restoration under diverse real-world degradation conditions, including blur, low-light, haze, rain, and snow. It provides a common benchmark for evaluating the restoration accuracy, robustness, and generalization capability of models across multiple degradation categories within a unified framework. The competition attracted 158 registered participants, and 20 teams were included in the final ranking after their submitted results were successfully reproduced and verified. This report provides a comprehensive analysis of the submitted solutions and corresponding results, highlighting recent advances in real-world all-in-one image restoration. The summarized methods and empirical findings reveal effective design strategies and establish an updated benchmark for future research in real-world low-level vision. △ Less

Submitted 23 July, 2026; originally announced July 2026.

Comments: ECCV 2026 Workshops; https://lowlevelcv.com/

Leveraging ECRAM for Edge Continual Learning

Authors: Nabila Tasnim, Haoran Liu, Qing Cao, Saugata Ghose

Abstract: Several edge computing platforms, such as autonomous vehicles and smart sensing devices, need to adapt to dynamic environments in real time by learning from new data in the field. Continual learning has emerged as a promising solution for edge training, by incorporating techniques that successfully combine a highly summarized version of previously trained data (to avoid catastrophic forgetting) wi… ▽ More Several edge computing platforms, such as autonomous vehicles and smart sensing devices, need to adapt to dynamic environments in real time by learning from new data in the field. Continual learning has emerged as a promising solution for edge training, by incorporating techniques that successfully combine a highly summarized version of previously trained data (to avoid catastrophic forgetting) with recently sensed data. However, as is the case with other ML algorithms, continual learning generates significant data movement between general-purpose CPUs/GPUs and memory, impacting the suitability of continual learning for edge platforms. In-memory computing (IMC; also known as processing-using-memory) can curtail this waste and make continual learning feasible at the edge, but it faces two unique challenges: (1) IMC architectures make use of noisy computation operations that significantly harm training accuracy; and (2) IMC architectures have poor and often incomplete support for resource-efficient training. To address these challenges, we propose CLASP (the Continual Learning Acceleration System Platform), which to our knowledge is the first end-to-end system with IMC acceleration for continual learning. The hardware and software of CLASP are co-designed to support a wide range of continual learning algorithms, through software-visible assembly-level instructions that can be incorporated without constraints into ML-based algorithms. CLASP is designed around a back-end-of-line (BEOL) compatible ECRAM device that we fabricate, which can overcome the challenges of IMC-based training using other emerging memory devices. We show that CLASP with ECRAM approaches the accuracy of in-GPU training, while delivering a speedup of 67x and energy savings of 132x for learning without forgetting and experience replay using MNIST. △ Less

Submitted 21 July, 2026; originally announced July 2026.

Comments: extended abstract of AICS 2025 poster

Ring-Zero: Scaling Zero RL to a Trillion Parameters for Emergent Reasoning

Authors: Xinyu Tang, Qianggang Cao, Yurou Liu, Yuliang Zhan, Xiaochong Lan, Yifan Li, Yuchen Yan, Han Peng, Zican Dong, Zhenduo Zhang, Tianshu Wang, Xinyu Kong, Zujie Wen, Wayne Xin Zhao, Zhiqiang Zhang, Jun Zhou

Abstract: Reinforcement learning with verifiable rewards without human-annotated data, often referred to as zero RL, has emerged as a powerful paradigm for eliciting chain-of-thought reasoning. However, due to computational constraints, existing studies are largely restricted to small models, leaving the training dynamics and emergent capabilities at a large scale unexplored. To meaningfully explore this fr… ▽ More Reinforcement learning with verifiable rewards without human-annotated data, often referred to as zero RL, has emerged as a powerful paradigm for eliciting chain-of-thought reasoning. However, due to computational constraints, existing studies are largely restricted to small models, leaving the training dynamics and emergent capabilities at a large scale unexplored. To meaningfully explore this frontier, we aim to elicit high-quality reasoning behaviors from the model. However, we find that naive scaling often suffers from poor readability, token redundancy, and a lack of adaptive reasoning depth. To address these challenges, we present a stable and efficient training pipeline, incorporating algorithmic and system optimizations such as clipped importance sampling, training-inference ratio correction, and mixed-precision control. Our experiments offer three key findings that validate the "bitter lesson" of scaling: (1) scaling to 1T parameters significantly enhances sample efficiency and performance ceilings; (2) the training process progresses sequentially through an initial discovery phase followed by a sharpening phase; and (3) the model spontaneously develops advanced cognitive behaviors, including anthropomorphism, structured formatting, self-verification, parallel reasoning, and context anxiety, rendering hand-crafted heuristics redundant. Evaluated on seven mathematical benchmarks, Ring-2.5-1T-Zero achieves competitive performance. Additionally, to assess CoT quality beyond final-answer correctness, we propose a structured evaluation framework across three dimensions: comprehensibility, reproducibility, and efficiency, where our model demonstrates clear advantages in producing structured and concise reasoning traces. By sharing our observed emergent phenomena, we hope to provide the community with deeper insights into scaling behaviors, particularly at the 1-trillion scale. △ Less

Submitted 15 July, 2026; v1 submitted 14 July, 2026; originally announced July 2026.

PixWorld: Unifying 3D Scene Generation and Reconstruction in Pixel Space

Authors: Sensen Gao, Zhaoqing Wang, Qihang Cao, Dongdong Yu, Changhu Wang, Jia-Wang Bian

Abstract: 3D reconstruction and generation are commonly tackled by separate paradigms: pixel-based regression for reconstruction, and latent diffusion for generation. Recent works attempt to unify them in latent space, but with notable drawbacks: the diffusion objective is defined on latent features rather than the underlying 3D representation, and both branches suffer from information loss introduced by la… ▽ More 3D reconstruction and generation are commonly tackled by separate paradigms: pixel-based regression for reconstruction, and latent diffusion for generation. Recent works attempt to unify them in latent space, but with notable drawbacks: the diffusion objective is defined on latent features rather than the underlying 3D representation, and both branches suffer from information loss introduced by latent encoding, while requiring a pretrained Variational Autoencoder (VAE) or Representation Autoencoder (RAE). In this paper, we reformulate these two tasks under a unified pixel-space diffusion paradigm and introduce PixWorld, a single model that jointly addresses 3D reconstruction and generation. By supervising diffusion directly on rendered images, PixWorld removes the above limitations and aligns optimization with 3D scene fidelity. Beyond photometric and perceptual supervision that operates at the 2D image level and lacks 3D geometric awareness, we further introduce a geometry perception loss that aligns rendered views with their ground truth in the geometry-aware feature space of a pretrained 3D foundation model, providing 3D structural supervision. PixWorld consistently outperforms prior latent-space generation methods and matches state-of-the-art reconstruction methods, demonstrating the superiority of a unified pixel-space approach. △ Less

Submitted 6 July, 2026; originally announced July 2026.

Comments: Project page: https://sensengao.github.io/PixWorld/

Uniqueness and Analytic Structures of Bosonic String Effective Amplitudes

Authors: Qu Cao, Fan Zhu

Abstract: We revisit the zero-transcendentality sector of bosonic string effective amplitudes with spin-1 external states, conjectured to correspond to a mass-deformed $(DF)^2$ theory, known as the $(DF)^2{+}\text{YM}$ theory. Imposing gauge invariance, locality, and cyclicity under minimal assumptions uniquely fixes a set of dimension-raising operators and leads to a recursive construction of amplitudes fr… ▽ More We revisit the zero-transcendentality sector of bosonic string effective amplitudes with spin-1 external states, conjectured to correspond to a mass-deformed $(DF)^2$ theory, known as the $(DF)^2{+}\text{YM}$ theory. Imposing gauge invariance, locality, and cyclicity under minimal assumptions uniquely fixes a set of dimension-raising operators and leads to a recursive construction of amplitudes from Yang-Mills amplitudes in the $α'{\to}0$ limit. At finite $α'$, certain derivative operators dressed with gauge invariant and $α'$-dependent factors, what we call $\textit{inverse operators}$, reconstruct the full bosonic string effective amplitudes, yielding compact expressions that universally factorize into tachyon-pole coefficients times Yang-Mills-Scalar amplitudes. This structure holds at arbitrary multiplicity and also extends to the amplitudes of the pure $(DF)^2$, $(DF)^2{+}φ^{3}$ and $(DF)^2{+}\text{YM}{+}φ^{3}$ theories. △ Less

Submitted 2 July, 2026; originally announced July 2026.

Comments: 12 pages, 5 figures and one ancillary file

Quantum Imaging via Kurtosis-Difference Weighted Covariance on 2D Camera

Authors: Zhe He, Yanli Shi, Hui Wu, Qun Cao, Weidong Zheng, Zheng Cui

Abstract: Camera-based quantum imaging detects spatially correlated photon pairs from spontaneous parametric down-conversion (SPDC). Conventional covariance methods typically require tens of thousands of frames to extract weak correlations from noise. While thick crystals can increase photon flux, they generate photon pairs from multiple emission positions within the crystal, producing multiple correlation… ▽ More Camera-based quantum imaging detects spatially correlated photon pairs from spontaneous parametric down-conversion (SPDC). Conventional covariance methods typically require tens of thousands of frames to extract weak correlations from noise. While thick crystals can increase photon flux, they generate photon pairs from multiple emission positions within the crystal, producing multiple correlation centers with complex pairing geometries. In addition, conventional covariance methods assume a single pre-selected correlation center and cannot fully exploit these distributed correlations. We demonstrate that kurtosis difference, a fourth-order statistic measuring tail similarity, effectively discriminates correlated pixel pairs even when correlation coefficients remain low. Weighting covariance by an exponential function of absolute kurtosis difference can select symmetric pixels while preserving true coincidences. This kurtosis weighting automatically identifies correlated pairs within a broad search region and accommodates multiple pairing geometries without requiring precise correlation center calibration. At 5000 frames, our method yields a contrast-to-noise ratio (CNR) exceeding 7, whereas standard covariance remains below 2. Compared with standard covariance, the method reduces the acquisition time by 40-fold and could enable practical quantum imaging in sparse correlated-photon regimes. △ Less

Submitted 29 June, 2026; originally announced June 2026.

Comments: 27 pages, 12 figures

Verifiable Auto-Formalization of Mathematics Using a Relaxed Natural Formal Language

Authors: Zhicheng Hui, Lihan Xie, Xingzhi Qi, Zhehao Li, Yingjun Lan, Qinxiang Cao

Abstract: Auto-formalization aims to translate informal mathematical content into formal languages that can be processed by theorem provers. However, directly targeting existing theorem provers requires LLMs to bridge a substantial representational gap between informal mathematical writing and formal proof languages. This gap also makes semantic consistency difficult to evaluate. We address these difficulti… ▽ More Auto-formalization aims to translate informal mathematical content into formal languages that can be processed by theorem provers. However, directly targeting existing theorem provers requires LLMs to bridge a substantial representational gap between informal mathematical writing and formal proof languages. This gap also makes semantic consistency difficult to evaluate. We address these difficulties by introducing a Relaxed Natural Formal Language (Relaxed NFL) as an intermediate target for auto-formalization. The Relaxed NFL is designed to remain close to informal mathematical writing: it preserves the usual structure of informal reasoning and allows partially specified expressions and propositions, without requiring their precise interpretation to be fixed at the auto-formalization stage. The remaining ambiguity and implicitness inherited from informal reasoning are resolved during a later elaboration stage, which transforms Relaxed NFL proofs into Core Natural Formal Language (Core NFL) proofs with formally defined semantics. The elaboration procedure combines rule-based transformations with LLM-generated heuristics, while maintaining verifiability through explicit constraints on each transformation step. The Core NFL is then used to generate proof gaps, namely verification conditions that must hold for the formalized proof to be correct. These gaps are discharged by LLM-generated proof scripts written in a domain-specific tactic language, which provides commands for invoking theorem libraries and domain-specific solvers implemented as part of our system. △ Less

Submitted 23 June, 2026; originally announced June 2026.

GUMP-Net: An interpretable model-data-driven intelligent algorithm for multi-class pelvic segmentation

Authors: Liheng Wang, Yinghui Zhang, Licheng Zhang, Hailin Xu, Qiyong Cao, Chong Chen

Abstract: Pelvic segmentation is one of the most important and fundamental research problems in precise and intelligent diagnosis and treatment, as well as surgical planning and navigation for pelvic fractures. By combining an improved geodesic active contour model with deep neural networks, we propose GUMP-Net, an interpretable model-data-driven intelligent algorithm for multi-class pelvic segmentation, in… ▽ More Pelvic segmentation is one of the most important and fundamental research problems in precise and intelligent diagnosis and treatment, as well as surgical planning and navigation for pelvic fractures. By combining an improved geodesic active contour model with deep neural networks, we propose GUMP-Net, an interpretable model-data-driven intelligent algorithm for multi-class pelvic segmentation, in which three network modules are designed to constitute the overall segmentation framework together: the object detection module for automatic level set initialization, the edge detector module for learning an anatomy-aware edge detector function and the iteration module for deep level set evolution. Leveraging the advantages of level set representation and deep learning, GUMP-Net shows more accurate, robust and consistent segmentation performance, especially in small training data situation, compared to the state-of-the-art methods. Extensive experiments on pelvic datasets demonstrate the rationality and effectiveness of the proposed algorithm. Further experiments extended to ankle dataset indicate broader applications to other anatomies. The proposed algorithm not only provides an efficient segmentation method for complex fracture reduction, but also gives an interpretable geometric perspective for understanding deep learning segmentation. △ Less

Submitted 17 June, 2026; originally announced June 2026.

Comments: 26 pages, 8 figures, 3 tables

MSC Class: 68U10; 94A08

Ling and Ring 2.6 Technical Report: Efficient and Instant Agentic Intelligence at Trillion-Parameter Scale

Authors: Ang Li, Ben Liu, Bin Han, Bin Hu, Bin Jing, Binbin Hu, Bing Li, Cai Chen, Caizhi Tang, Changxin Tian, Chao Huang, Chao Zhang, Chen Liang, Chen Qian, Chengfu Tang, Chengyao Wen, Chilin Fu, Chunwei Wu, Cong Zhang, Cunyin Peng, Daixin Wang, Dalong Zhang, Deng Zhao, Dingnan Jin, Dingyuan Zhu , et al. (193 additional authors not shown)

Abstract: Efficient and scalable agentic intelligence requires models that can deliver both low-latency responses and strong reasoning capabilities while remaining practical to train, serve, and deploy. In this report, we present Ling-2.6 and Ring-2.6, a family of models designed to address this challenge at scale. Ling-2.6 is optimized for instant response generation and high capability per output token, w… ▽ More Efficient and scalable agentic intelligence requires models that can deliver both low-latency responses and strong reasoning capabilities while remaining practical to train, serve, and deploy. In this report, we present Ling-2.6 and Ring-2.6, a family of models designed to address this challenge at scale. Ling-2.6 is optimized for instant response generation and high capability per output token, whereas Ring-2.6 is tailored for deeper reasoning and more advanced agentic workflows. Instead of training from scratch, we upgrade the Ling-2.0 base model through architectural migration pre-training and large-scale post-training. This upgrade is guided by a unified co-design of model architecture, optimization objectives, serving systems, and agent training environments, enabling improvements in both model capability and deployment efficiency. At the architectural level, we introduce a hybrid linear attention design that integrates Lightning Attention with MLA, improving the efficiency of long-context training and decoding. To further enhance token efficiency, we optimize capability per output token through Evolutionary Chain-of-Thought, Linguistic Unit Policy Optimization, bidirectional preference alignment, and shortest-correct-response distillation. For agentic capabilities, we propose KPop, a reinforcement learning framework designed to support stable training of Ring-2.6-1T on large-scale environment-grounded data. KPop improves training efficiency through asynchronous scheduling across coding, search, tool use, and workflow execution, enabling scalable learning from complex agent-environment interactions. Together, Ling-2.6 and Ring-2.6 provide a practical pathway toward efficient, scalable, and open agentic systems. We open-source all checkpoints in the 2.6 family to support further research and development in practical agentic intelligence. △ Less

Submitted 12 June, 2026; originally announced June 2026.

MIDSim: Simulating Multi-Channel Information Diffusion in Social Media with LLM-Powered Multi-Agent System

Authors: Lexi Liu, Qi Cao, Yuanhao Liu, Huawei Shen, Xueqi Cheng

Abstract: Information diffusion in social media shapes public opinion and collective behavior, making its modeling and simulation an important research problem. Existing studies have investigated information diffusion through epidemic-based, cascade-based, and point process models. However, they predominantly focus on diffusion through social links, overlooking other diffusion channels enabled by platform a… ▽ More Information diffusion in social media shapes public opinion and collective behavior, making its modeling and simulation an important research problem. Existing studies have investigated information diffusion through epidemic-based, cascade-based, and point process models. However, they predominantly focus on diffusion through social links, overlooking other diffusion channels enabled by platform algorithms (e.g., recommender systems) and failing to capture user behavioral complexity. To address these limitations, we propose an LLM-powered multi-agent system for simulating multi-channel information diffusion, where large language models instantiate personalized user agents and the diffusion process jointly models social and algorithmic exposure streams. We further construct three real-world diffusion dataset spanning Sina Weibo, RedNote, and Twitter, containing diffusion records, user profiles, historical posts, and social relationships. Experimental results on real diffusion events show that our proposed framework realistically simulate macro diffusion phenomenon and generate diverse comment content, significantly outperforming baselines. △ Less

Submitted 15 June, 2026; v1 submitted 11 June, 2026; originally announced June 2026.

SearchSwarm: Towards Delegation Intelligence in Agentic LLMs for Long-Horizon Deep Research

Authors: Xiaochong Lan, Quan Chen, Kun Tao, Xinyu Tang, Tianshu Wang, Qianggang Cao, Xinyu Kong, Zujie Wen, Zhiqiang Zhang, Jun Zhou

Abstract: Large language models are increasingly expected to handle complex, long-horizon real-world tasks whose context demands can grow without bound, yet model context windows remain inherently finite. Recent work explores a paradigm where a main agent decomposes tasks and dispatches subtasks to subagents, which execute and return only summarized results, conserving the main agent's context budget. Howev… ▽ More Large language models are increasingly expected to handle complex, long-horizon real-world tasks whose context demands can grow without bound, yet model context windows remain inherently finite. Recent work explores a paradigm where a main agent decomposes tasks and dispatches subtasks to subagents, which execute and return only summarized results, conserving the main agent's context budget. However, performing this well requires delegation intelligence: the ability to decompose complex tasks, determine when and what to delegate, and integrate returned results into the ongoing workflow. Training data for this capability is scarce in naturally occurring text, and to our knowledge, how to synthesize such data and train models to acquire this capability remains largely unexplored in the open-source community. To bridge this gap, we present a preliminary exploration targeting deep research, a representative long-horizon agent task. Specifically, we design a harness that guides the model toward high-quality task decomposition and delegation, while constraining subagents to return results properly to support the main agent's workflow. The harness-guided trajectories naturally encode correct delegation decisions, which we use as supervised fine-tuning data to internalize delegation intelligence into model weights. Our resulting model, SearchSwarm-30B-A3B, achieves 68.1 on BrowseComp and 73.3 on BrowseComp-ZH, the best results among all models of comparable scale. We will release our harness, model weights, and training data to facilitate future research. △ Less

Submitted 9 August, 2026; v1 submitted 8 June, 2026; originally announced June 2026.

Activation Steering Induces Emergent Misalignment: A More Comprehensive Evaluation

Authors: Qi Cao, Jian Lou, Meiting Liu, Wenjie Feng, Dan Li, See-Kiong Ng, Anh Tuan Luu

Abstract: Activation steering has emerged as a popular inference-time technique for modulating the behavior of large language models (LLMs). By constructing a steering vector from examples of a target behavior and injecting it into intermediate activations during inference, activation steering enables flexible behavioral control while avoiding the permanent parameter updates required by finetuning. Meanwhil… ▽ More Activation steering has emerged as a popular inference-time technique for modulating the behavior of large language models (LLMs). By constructing a steering vector from examples of a target behavior and injecting it into intermediate activations during inference, activation steering enables flexible behavioral control while avoiding the permanent parameter updates required by finetuning. Meanwhile, recent work has identified emergent misalignment (EM) as a significant safety concern, wherein models finetuned on unsafe examples from a narrow task may unexpectedly generalize to broadly unsafe behavior on unrelated tasks. Although finetuning-induced EM has been extensively studied, whether activation steering can induce EM remains comparatively under-explored, despite its increasing use as a model-control technique. In this paper, we present a comprehensive study of activation-steering-induced emergent misalignment, substantially expanding the evaluation scope beyond existing pioneering work. First, we show that activation steering can induce broad misalignment, even in the recent Qwen-3.5 series. Moreover, activation-steered models produce harmful responses with stronger semantic relevance and higher coherence than their finetuned counterparts, making the resulting misalignment potentially more harmful. Second, we characterize properties of AS-induced EM by analyzing key steering-specific factors, including steering magnitude, the low-rank structure of the steering subspace, and the number of epochs during steering-vector construction. Third, we evaluate the robustness and sensitivity of AS-induced EM across diverse model families, model scales, target tasks, and intervention layers. Our findings reveal activation steering as a significant yet under-examined source of emergent misalignment and provide an activation-space perspective for understanding the mechanisms and safety risks of EM. △ Less

Submitted 7 June, 2026; originally announced June 2026.

ResearchClawBench: A Benchmark for End-to-End Autonomous Scientific Research

Authors: Wanghan Xu, Shuo Li, Tianlin Ye, Qinglong Cao, Yixin Chen, Hengjian Gao, Yiheng Wang, Qi Li, Kun Li, Sheng Xu, Shengdu Chai, Fangchen Yu, Xiangyu Zhao, Zhangrui Zhao, Weijie Ma, Zijie Guo, Koutian Wu, Haoyu Zhou, Haoxiang Yin, Lixue Cheng, Chaofan Hu, Haoxuan Li, Lu Mi, Xuxuan Xie, Yifan Zhou , et al. (26 additional authors not shown)

Abstract: AI coding agents are increasingly used for scientific work, but their end-to-end autonomous research capability remains difficult to verify. We present ResearchClawBench, a benchmark for evaluating autonomous scientific research across 40 tasks from 10 scientific domains. Each task is grounded in a real published paper, provides related literature and raw data, and hides the target paper during ev… ▽ More AI coding agents are increasingly used for scientific work, but their end-to-end autonomous research capability remains difficult to verify. We present ResearchClawBench, a benchmark for evaluating autonomous scientific research across 40 tasks from 10 scientific domains. Each task is grounded in a real published paper, provides related literature and raw data, and hides the target paper during evaluation. Expert-curated multimodal rubrics decompose the target scientific artifacts into weighted criteria, enabling evaluation of target-paper-level re-discovery while leaving room for new discovery. We evaluate seven autonomous research (auto-research) agents under a unified protocol and seventeen native LLMs through the lightweight ResearchHarness. Current systems remain far from reliable re-discovery: the strongest autonomous agent, Claude Code, averages 21.5, and the strongest ResearchHarness LLM, Claude-Opus-4.7, averages 20.7, with an LLM frontier mean of only 26.5. Error analysis shows that failures concentrate in experimental protocol mismatch, evidence mismatch, and missing scientific core. ResearchClawBench provides a reproducible evaluation frontier for measuring progress toward autonomous scientific research. △ Less

Submitted 2 July, 2026; v1 submitted 28 May, 2026; originally announced June 2026.

Fermiology and the Candidate Chiral Superconductor in Rhombohedral Tetralayer Graphene

Authors: Sandesh S. Kalantre, Ben H. Alexander, Julian May-Mann, Jonah Herzog-Arbeitman, Marisa Hocking, Qingrui Cao, Kenji Watanabe, Takashi Taniguchi, David Goldhaber-Gordon, Andrew J. Mannix, Trithep Devakul, Yves H. Kwan, Daniel E. Parker, Aaron Sharpe

Abstract: Chiral superconductivity, in which the phase of the superconducting order parameter winds in momentum space, has long been sought for its close link to topological superconductivity. Recent work reported a superconductor in rhombohedral multilayer graphene emerging from a time-reversal symmetry broken normal state, suggesting that it could be a chiral superconductor. However, the possibility of ch… ▽ More Chiral superconductivity, in which the phase of the superconducting order parameter winds in momentum space, has long been sought for its close link to topological superconductivity. Recent work reported a superconductor in rhombohedral multilayer graphene emerging from a time-reversal symmetry broken normal state, suggesting that it could be a chiral superconductor. However, the possibility of chirality depends on the symmetry and structure of the normal-state Fermi surface, which have not been directly measured. Here we measure quantum oscillations in rhombohedral tetralayer graphene over a broad range of the phase diagram, including the superconducting region. At densities well above the onset of superconductivity, we reproduce previously-reported oscillations consistent with a spin- and valley-polarized quarter metal with a single simply-connected Fermi pocket. As the carrier density is reduced, we find a transition to a complex "multitone" state that persists through the superconducting region. This state's spectrum of quantum oscillations is incompatible with a simply-connected quarter metal. The next-simplest candidate normal states suggested by our microscopic modeling (fully-polarized annular, nematic, and three-pocket states) are inconsistent with our measurements, albeit difficult to rule out entirely. The normal state is thus seen to be richer than previously envisaged, reshaping the search for the superconducting mechanism and the possible chirality of the pairing channel. △ Less

Submitted 3 June, 2026; originally announced June 2026.

Comments: 12+60 pages, 4+52 figures

SCI-PRM: A Tool Aware Process Reward Model for Scientific Reasoning Verification

Authors: Xiangyu Zhao, Henry Hengyuan Zhao, Yiheng Wang, Wanghan Xu, Yuhao Zhou, Qinglong Cao, Zhiwang Zhou, Lei Bai, Wenlong Zhang, Xiao-Ming Wu

Abstract: While Process Reward Models (PRMs) have achieved remarkable success in mathematical reasoning, their application in complex scientific domains-such as biology, chemistry, and physics remains largely unexplored. Scientific problems demand not only logical rigor but also factual consistency and the precise usage of domain-specific tools, areas where current models often suffer from hallucinations an… ▽ More While Process Reward Models (PRMs) have achieved remarkable success in mathematical reasoning, their application in complex scientific domains-such as biology, chemistry, and physics remains largely unexplored. Scientific problems demand not only logical rigor but also factual consistency and the precise usage of domain-specific tools, areas where current models often suffer from hallucinations and lack of verification. In this paper, we first construct SCIPRM70K, a large-scale dataset featuring Chain-of-Tool trajectories that explicitly interleave reasoning with the execution of scientific tools. Building upon this, we train an efficient reward model called Sci-PRM to provide fine-grained supervision on tool selection, execution accuracy, and result interpretation at each step in one inference. Experiments demonstrate that Sci-PRM significantly enhances foundation models in two key aspects: (1) it enables effective test-time scaling via Best-of-N selection; and (2) when integrated into Reinforcement Learning, it serves as a dense reward signal that mitigates the critical issue of advantage disappearance, allowing the model to break through existing performance ceilings. △ Less

Submitted 22 June, 2026; v1 submitted 3 June, 2026; originally announced June 2026.

Comments: Accepted by KDD 2026 AI4Science Track

Clustered Self-Assessment: A Simple yet Effective Method for Uncertainty Quantification in Large Language Models

Authors: Qi Cao, Takeshi Kojima, Andrew Gambardella, Helinyi Peng, Yutaka Matsuo, Yusuke Iwasawa

Abstract: Large language models (LLMs) demonstrate remarkable performance across diverse tasks, but they often generate responses that appear plausible while being factually incorrect. This problem is compounded by the lack of explicit uncertainty estimates, which makes it difficult for users to judge the reliability of model outputs. Existing uncertainty quantification methods typically rely on indirect si… ▽ More Large language models (LLMs) demonstrate remarkable performance across diverse tasks, but they often generate responses that appear plausible while being factually incorrect. This problem is compounded by the lack of explicit uncertainty estimates, which makes it difficult for users to judge the reliability of model outputs. Existing uncertainty quantification methods typically rely on indirect signals, such as entropy across sampled generations. These signals can be difficult to interpret and do not fully leverage the model's ability to assess its own uncertainty. We propose a simple yet effective self-assessment method for uncertainty quantification in LLMs. Our approach groups sampled generations into semantically distinct clusters, converts them into answer options in a structured multiple-choice question, and uses the probability assigned by the LLM to each option as a confidence estimate. Experiments across multiple models and datasets show that our method consistently outperforms baseline approaches. Notably, it achieves competitive performance with as few as two additional samples, demonstrating both its effectiveness and efficiency. △ Less

Submitted 2 June, 2026; originally announced June 2026.

Comments: Findings of ACL 2026

ATLAS: Agentic Test-time Learning-to-Allocate Scaling

Authors: Peijia Qin, Qi Cao, Pengtao Xie

Abstract: Test-time scaling has become a major way to improve large language model reasoning, but its orchestration has remained designer-engineered: a fixed sample budget, a fixed refinement loop, a fixed scoring rule, or a fixed search policy decides how compute is spent, leaving the model in charge of solving but not of orchestration. We introduce ATLAS, an agentic test-time scaling framework in which an… ▽ More Test-time scaling has become a major way to improve large language model reasoning, but its orchestration has remained designer-engineered: a fixed sample budget, a fixed refinement loop, a fixed scoring rule, or a fixed search policy decides how compute is spent, leaving the model in charge of solving but not of orchestration. We introduce ATLAS, an agentic test-time scaling framework in which an LLM orchestrator owns the control loop end-to-end. Through a single action, explore, which dispatches a fresh independent solver on the original problem, the orchestrator decides whether to gather more evidence, when to stop, and how to synthesize the final answer; the action space is extensible, with each explore call optionally specifying solver, reasoning effort, or prompting strategy. We evaluate ATLAS on four benchmarks covering scientific question answering, code generation, and multimodal reasoning under a Claude Sonnet 4.6 backbone, where it reaches 56.00% on HLE-Verified, 82.29% on LiveCodeBench, 85.75% on GPQA-Diamond, and 23.71% on BabyVision while using far fewer API calls than fixed-workflow baselines. A multi-model extension, ATLAS-MM, that exposes solver choice as an additional action dimension further improves HLE-Verified to 60.00% and LiveCodeBench to 85.63%, with consistent gains on GPQA-Diamond and BabyVision. Ablations replacing the orchestrator's direct synthesis with a separate integrator degrade or fail to improve accuracy on three of four benchmarks, consistent with the role of stateful evidence management in producing the gains. △ Less

Submitted 1 June, 2026; originally announced June 2026.

Send a SCOUT First: Pre-hoc Reasoning for Adaptive Detector Allocation in Prompt-Injection Defense

Authors: Shuhao Zhang, Jiarui Li, Qi Cao, Ruiyi Zhang, Pengtao Xie

Abstract: Prompt-injection detectors are heterogeneous: each is strong on a different slice of attacks, and none is always reliable. Yet existing systems still treat detection as a fixed single-detector pipeline, committing every request to one detector's blind spots. We reframe defense as detector allocation: given a heterogeneous pool, decide per request which detectors to run and whether to escalate to a… ▽ More Prompt-injection detectors are heterogeneous: each is strong on a different slice of attacks, and none is always reliable. Yet existing systems still treat detection as a fixed single-detector pipeline, committing every request to one detector's blind spots. We reframe defense as detector allocation: given a heterogeneous pool, decide per request which detectors to run and whether to escalate to an LLM judge. Our framework SCOUT (Scalable and Controllable Outcome-prediction for Uncertainty-aware Triage) makes this decision dynamic by predicting each detector's per-sample reliability and latency from how it behaved on similar past inputs, and exposes a single safety-utility threshold to the operator (where utility bundles benign-pass rate and wall-clock). To evaluate this setting, we build SCOUT-450, a benchmark that captures the structurally complex, agent-facing injections that older prompt-injection sets under-represent. On SCOUT-450, a safety-oriented operating point reduces attack-success rate by 46% and total wall-clock by 40% relative to an always-on GPT-4o judge, at a 5.1-point benign-utility drop. SCOUT also transfers to three external benchmarks (BIPIA, IPI, and IHEval), improving the safety-utility frontier. △ Less

Submitted 14 June, 2026; v1 submitted 29 May, 2026; originally announced May 2026.

Comments: We propose SCOUT, a detector allocation framework that predicts each detector's accuracy and latency on a given input before running it, letting operators control the safety-utility trade-off with a single threshold and route to an LLM judge only when needed

Visualizing orbital magnetism in electron doped rhombohedral multilayer graphene

Authors: Owen I. Sheekey, Trevor B. Arp, Benjamin A. Foutty, Ruoxi Zhang, Tixuan Tan, Ludwig F. W. Holleis, Yi Guo, Sandesh S. Kalantre, Canxun Zhang, Mark Zakharyan, David Gong, Aidan Keough, Youngjoon Choi, Ysun Choi, Siyuan Xu, Tian Xie, Ben Hodder Alexander, Marisa Hocking, Qingrui Cao, Martin E. Huber, Takashi Taniguchi, Kenji Watanabe, Chenhao Jin, Etienne Lantagne-Hurtubise, Aaron Sharpe , et al. (2 additional authors not shown)

Abstract: Electron doped rhombohedral multilayer graphene at high displacement field features an exceptionally flat band minimum with near-ideal quantum geometry. Experiments in this regime observe the formation of a 'quarter metal,' in which the electron liquid condenses into a single spin- and valley flavor. Remarkably, recent experiments have found a zero resistance state in the same region of the densit… ▽ More Electron doped rhombohedral multilayer graphene at high displacement field features an exceptionally flat band minimum with near-ideal quantum geometry. Experiments in this regime observe the formation of a 'quarter metal,' in which the electron liquid condenses into a single spin- and valley flavor. Remarkably, recent experiments have found a zero resistance state in the same region of the density- and displacement-field-tuned parameter space, attributed to the formation of a chiral superconductor from an orbitally ferromagnetic normal state. Here, we use nanoSQUID-on-tip magnetometry to map the orbital magnetization of electron-doped rhombohedral graphene devices ranging in thickness between 3 and 15 layers. Magnetization within the quarter metal phases peaks at finite density, consistent with concentration of the Berry curvature in a finite-momentum 'ring of fire'. Correlating transport and local magnetometry data in a superconducting tetralayer sample reveals a finite orbital ferromagnetic moment, providing direct evidence of valley polarization in the superconducting ground state. We further show that widely observed stochastic switching of the resistivity in both metallic and superconducting regimes arises from a density-tuned sign change in the valley-resolved total magnetic moment. This leads to the formation of metastable magnetic domains under typical gate control sequences and can also be harnessed for electric-field controlled switching of the magnetization across the entire device. Finally, high resolution measurements of the magnetization across a superconducting transition allow us to put an upper bound on the 'condensation magnetization' of 0.1 Bohr magneton per carrier, placing a strong quantitative restriction on theoretical models for ferromagnetic superconductivity. △ Less

Submitted 4 August, 2026; v1 submitted 28 May, 2026; originally announced May 2026.

MINDGAMES: A Live Arena for Evaluating Social and Strategic Reasoning in Multi-Agent LLMs

Authors: Kevin Wang, Anna Thöni, Benjamin Kempinski, Bobby Cheng, Jianzhu Yao, Benjamin Finch, Leon Guertler, Viraj Nadkarni, Yihan Jiang, Aliaksei Korshuk, Alexander Buyantuev, Ilya Makarov, Siyuan Wu, Yu-Chi Cheng, Yan-Ru Ju, Ti-Rong Wu, I-Hsuan Chu, Yu-Yu Yang, I-Chen Wu, Yitian Huang, Qinlu Cao, Yiheng Sun, Yuhong Dai, Hongkun Yao, Jingxuan Fu , et al. (28 additional authors not shown)

Abstract: Large language models (LLMs) are increasingly deployed as interactive agents, yet their capacity for social and strategic reasoning over extended interaction remains poorly understood. Existing evaluations rely on static vignettes or single-game benchmarks that cannot capture the sustained, multi-faceted reasoning that real-world multi-agent settings demand. We introduce Mindgames, a multi-game ar… ▽ More Large language models (LLMs) are increasingly deployed as interactive agents, yet their capacity for social and strategic reasoning over extended interaction remains poorly understood. Existing evaluations rely on static vignettes or single-game benchmarks that cannot capture the sustained, multi-faceted reasoning that real-world multi-agent settings demand. We introduce Mindgames, a multi-game arena and evaluation platform for LLM agents that operationalizes complementary reasoning demands relevant to ``theory of mind'': belief attribution under hidden information, opponent modeling through repeated strategic interaction, cooperative inference under knowledge asymmetries, and sustained deception in social deduction. Built on TextArena, Mindgames provides a unified interaction interface, TrueSkill-based rating, and full trajectory logging across four game environments. We instantiate Mindgames through a 2025 competition cycle hosted at a major AI conference, which assessed 944 submitted agents from 76 teams across four games: Colonel Blotto, Iterated Prisoner's Dilemma, Codenames, and Secret Mafia. Our analysis surfaces both agent-level and evaluation-level limitations: brittle rule adherence remains a major bottleneck, top-performing systems repeatedly rely on explicit structural scaffolding, and leaderboard validity differs sharply across environments. In particular, failure-heavy environments can reward robustness to opponent errors as much as strategic ability, with Secret Mafia exhibiting a pronounced error-survival confound in this cycle. We release a dataset of 29,571 multi-agent games with turn-level observations, actions, and rewards, together with MG-Ref, a deterministic offline tournament protocol that scores new agents against a frozen reference pool of top-ranked, low-error Stage~II submissions under the same error-attribution lens used in this analysis. △ Less

Submitted 28 May, 2026; originally announced May 2026.

AIBuildAI-2: A Knowledge-Enhanced Agent for Automatically Building AI Models

Authors: Ruiyi Zhang, Peijia Qin, Qi Cao, Li Zhang, Pengtao Xie

Abstract: AI models underpin data-centric applications from image and text processing to scientific discovery in biology, physics, and chemistry. Yet developing them remains heavily manual, requiring practitioners to design architectures, build training pipelines, and iteratively refine solutions, making it challenging for natural scientists without specialized AI engineering expertise to build the high-per… ▽ More AI models underpin data-centric applications from image and text processing to scientific discovery in biology, physics, and chemistry. Yet developing them remains heavily manual, requiring practitioners to design architectures, build training pipelines, and iteratively refine solutions, making it challenging for natural scientists without specialized AI engineering expertise to build the high-performing models their research demands. To reduce this burden and broaden access to AI for scientific discovery, agents that automatically build AI models have been proposed. However, the performance of these agents is largely limited by the parametric knowledge of their underlying large language models, which is static, often outdated, and sparse on practical AI model engineering know-how. To address this limitation, we introduce AIBuildAI-2, a knowledge-enhanced agent with an external, evolving knowledge system for automatically building AI models. The knowledge system of AIBuildAI-2 is hierarchical, organizing curated AI development knowledge into high-level knowledge instructions over topical categories and low-level knowledge documents under each category, from which the agent dynamically loads only the context relevant to its current state and the AI task being solved, grounding each design and implementation decision in concrete, externally verifiable expertise. The system is initialized by collecting and cleaning AI-development-related documents from the web and organizing them into the corresponding categories, and continually evolves from the agent's own experience by distilling each completed run on an AI task into structured takeaways that are written back into the knowledge system. AIBuildAI-2 achieves state-of-the-art results, ranking first on MLE-Bench with a 70.7% medal rate and placing in the top 6.6% among 4,370 human-expert teams in a heart disease prediction competition. △ Less

Submitted 26 May, 2026; originally announced May 2026.

Symmetry Breaking as Quantum Gate: Entropy and Weak Mixing Angle

Authors: Qing-Hong Cao, Yandong Liu, Haotian Qi, Hao Zhang, Haoran Zhao

Abstract: We establish a correspondence between two independent entropic probes -- the variation of Rényi mutual information (RMI) across the electroweak symmetry breaking (EWSB) transition and the stabilizer Rényi entropy (SRE) -- in tree-level $2\to 2$ elastic scatterings. After angular averaging, the RMI (helicity basis) and the SRE (fixed beam basis) exhibit identical dependence on $\sin^2θ_W$ within ea… ▽ More We establish a correspondence between two independent entropic probes -- the variation of Rényi mutual information (RMI) across the electroweak symmetry breaking (EWSB) transition and the stabilizer Rényi entropy (SRE) -- in tree-level $2\to 2$ elastic scatterings. After angular averaging, the RMI (helicity basis) and the SRE (fixed beam basis) exhibit identical dependence on $\sin^2θ_W$ within each neutral-current channel. We trace this correspondence to a common physical origin that it's the Yukawa mass insertion acts as a $-\mathrm{i}Y$ quantum gate in chirality space. Minimizing entropies across all processes yields $\sin^2θ_W$ values matching purely axial vector-like couplings in $Z$ boson exchanged channel. △ Less

Submitted 21 May, 2026; originally announced May 2026.

Comments: 6 pages, 7 figures

FishBack: Pullback Fisher Geometry for Optimal Activation Steering in Transformers

Authors: Sihan Wang, Jiayi Zhao, Qingyan Cao, Hongbo Yao, Lin Shu

Abstract: Activation steering has emerged as a lightweight approach for modifying language model behavior without parameter updates, yet existing methods remain brittle: unstable across layers and prone to disturbing behavior unrelated to the target concept. We trace these failures to a hidden assumption shared by widely-used methods such as CAA, ActAdd, and ITI: that the intermediate activation space is Eu… ▽ More Activation steering has emerged as a lightweight approach for modifying language model behavior without parameter updates, yet existing methods remain brittle: unstable across layers and prone to disturbing behavior unrelated to the target concept. We trace these failures to a hidden assumption shared by widely-used methods such as CAA, ActAdd, and ITI: that the intermediate activation space is Euclidean. We show this assumption is fundamentally flawed. The metric that actually governs how a hidden-state perturbation changes the output is the Fisher information metric of the softmax layer, pulled back to the intermediate layer through the Jacobian of the intervening layers. From it we derive a closed-form steering direction, applied to a hidden state at an intermediate layer, that reaches a target concept change with the least non-target distortion. The framework is sharpest in the early and middle intermediate layers, where the metric is strongly non-Euclidean and geometric correction matters most. We evaluate it on three verb-morphology concepts: third-person-singular, progressive, and past-tense inflection, following standard counterfactual-concept evaluation. On GPT-2 Small, this non-Euclidean geometry is borne out empirically, and our method lowers off-target KL divergence by median factors of 1.4--6.5x against individual steering baselines. On Llama-3-8B and Qwen3-8B, it lowers off-target KL by median factors of 1.8--3.6x against individual baselines at the early and middle layers. These results show that geometric correction retains its advantage on larger models with more complex internal structure. △ Less

Submitted 18 August, 2026; v1 submitted 16 May, 2026; originally announced May 2026.

Comments: Preprint. 22 pages, 6 figures, 14 tables

Measurement and Control of the Complex Berry Phase in a Quantum System

Authors: Pratik J. Barge, Qian Cao, Niklas Hörnedal, Aurélia Chenu, Kater W. Murch

Abstract: The Berry phase is a geometric phase acquired during adiabatic evolution over a closed loop in parameter space. It plays an essential role in geometric quantum gates and other phase-based protocols. In non-Hermitian systems, the Berry phase is complex, introducing fundamentally new geometric effects, including state amplification. In this work, we report experimental measurement of both the real a… ▽ More The Berry phase is a geometric phase acquired during adiabatic evolution over a closed loop in parameter space. It plays an essential role in geometric quantum gates and other phase-based protocols. In non-Hermitian systems, the Berry phase is complex, introducing fundamentally new geometric effects, including state amplification. In this work, we report experimental measurement of both the real and imaginary components of a Berry phase in a fully quantum system using a superconducting transmon circuit with engineered dissipation. We also demonstrate the path-dependent effects of the imaginary part on the dissipation and its utility in the implementation of non-unitary quantum control. These findings establish a clear geometric distinction between the real and imaginary components of the Berry phase and experimentally confirm the unique adiabatic behavior of non-Hermitian quantum systems. △ Less

Submitted 15 May, 2026; originally announced May 2026.

Comments: 10 pages, 8 figures

LLMs Know When They Know, but Do Not Act on It: A Metacognitive Harness for Test-time Scaling

Authors: Qi Cao, Yufan Wang, Peijia Qin, Shuhao Zhang, Pengtao Xie

Abstract: Large language models (LLMs) often expose useful signals of self-monitoring: before solving a problem, they can estimate whether they are likely to succeed, and after solving it, they can judge whether their answer is likely to be correct. However, these signals are typically measured or elicited in isolation, rather than used to control inference. In this work, we ask whether LLMs possess latent… ▽ More Large language models (LLMs) often expose useful signals of self-monitoring: before solving a problem, they can estimate whether they are likely to succeed, and after solving it, they can judge whether their answer is likely to be correct. However, these signals are typically measured or elicited in isolation, rather than used to control inference. In this work, we ask whether LLMs possess latent metacognitive ability that can be turned into effective test-time control. Inspired by the Nelson--Narens theory from cognitive psychology, we propose a metacognitive harness that separates monitoring from reasoning. For each problem, the model first reports a pre-solve feeling-of-knowing (FOK) signal; after each solve attempt, it reports a post-solve judgment-of-learning (JOL) signal. Rather than treating these signals as passive confidence estimates, the harness turns them into an explicit control interface for reasoning: it decides when to trust the current solution, when to retry with compact metacognitive feedback, and when to pass multiple attempts to a final aggregator. Across text, code, and multimodal reasoning benchmarks, our harness substantially improves a fixed Claude Sonnet-4.6 base model without parameter updates or benchmark-specific fine-tuning. On the evaluated public benchmark snapshots, it raises pooled accuracy from 48.3 to 56.9 and exceeds the strongest listed leaderboard entries on the three primary evaluation settings: HLE-Verified, LiveCodeBench v6, and R-Bench-V. These results suggest that strong LLMs may already possess useful metacognitive ability, but require an explicit control harness to act on it during reasoning. △ Less

Submitted 13 May, 2026; originally announced May 2026.

Qwen-Scope: Turning Sparse Features into Development Tools for Large Language Models

Authors: Boyi Deng, Xu Wang, Yaoning Wang, Yu Wan, Yubo Ma, Baosong Yang, Haoran Wei, Jialong Tang, Huan Lin, Ruize Gao, Tianhao Li, Qian Cao, Xuancheng Ren, Xiaodong Deng, An Yang, Fei Huang, Dayiheng Liu, Jingren Zhou

Abstract: Large language models have achieved remarkable capabilities across diverse tasks, yet their internal decision-making processes remain largely opaque, limiting our ability to inspect, control, and systematically improve them. This opacity motivates a growing body of research in mechanistic interpretability, with sparse autoencoders (SAEs) emerging as one of the most promising tools for decomposing… ▽ More Large language models have achieved remarkable capabilities across diverse tasks, yet their internal decision-making processes remain largely opaque, limiting our ability to inspect, control, and systematically improve them. This opacity motivates a growing body of research in mechanistic interpretability, with sparse autoencoders (SAEs) emerging as one of the most promising tools for decomposing model activations into sparse, interpretable feature representations. We introduce Qwen-Scope, an open-source suite of SAEs built on the Qwen model family, comprising 14 groups of SAEs across 7 model variants from the Qwen3 and Qwen3.5 series, covering both dense and mixture-of-expert architectures. Built on top of these SAEs, we show that SAEs can go beyond post-hoc analysis to serve as practical interfaces for model development along four directions: (i) inference-time steering, where SAE feature directions control language, concepts, and preferences without modifying model weights; (ii) evaluation analysis, where activated SAE features provide a representation-level proxy for benchmark redundancy and capability coverage; (iii) data-centric workflows, where SAE features support multilingual toxicity classification and safety-oriented data synthesis; and (iv) post-training optimization, where SAE-derived signals are incorporated into supervised fine-tuning and reinforcement learning objectives to mitigate undesirable behaviors such as code-switching and repetition. Together, these results demonstrate that SAEs can serve not only as post-hoc analysis tools, but also as reusable representation-level interfaces for diagnosing, controlling, evaluating, and improving large language models. By open-sourcing Qwen-Scope, we aim to support mechanistic research and accelerate practical workflows that connect model internals to downstream behavior. △ Less

Submitted 12 May, 2026; originally announced May 2026.

Why Do Aligned LLMs Remain Jailbreakable: Refusal-Escape Directions, Operator-Level Sources, and Safety-Utility Trade-off

Authors: Yu Chen, Yuanhao Liu, Qi Cao

Abstract: Aligned large language models (LLMs) remain vulnerable to jailbreak attacks. Recent mechanistic studies have identified latent features and representation shifts associated with jailbreak success, but they leave a more fundamental question open: why do aligned LLMs remain jailbreakable, and what structural vulnerabilities in the model make this possible? We study this question through a continuous… ▽ More Aligned large language models (LLMs) remain vulnerable to jailbreak attacks. Recent mechanistic studies have identified latent features and representation shifts associated with jailbreak success, but they leave a more fundamental question open: why do aligned LLMs remain jailbreakable, and what structural vulnerabilities in the model make this possible? We study this question through a continuous input-transformation view. Our theoretical finding is that aligned models can still exhibit Refusal-Escape Directions (RED): local perturbation directions around a harmful input that shift the model's behavior from refusal to answering while preserving the model's harmful-semantics interpretation. From this perspective, a jailbreak is not only a successful discrete prompt construction, but can also be understood as a refusal-to-answer behavior transition induced by continuously perturbing a harmful input along RED. We then prove that RED can be exactly decomposed into contributions from operator-level sources across the model's operator structure, and identify normalization, residual-wiring, and terminal sources as analytically constrained operator-level sources. To eliminate RED, the shared expressive modules -- self-attention and MLP -- must eliminate the contributions from these analytically constrained sources while preserving the mechanisms that support benign responses. These competing requirements give rise to a conditional safety-utility trade-off. Experiments across multiple models and attack methods empirically analyze RED from two complementary perspectives and show that added token dimensions can expose RED, while successful jailbreaks exhibit refusal-to-answer shifts largely aligned with terminal-source contributions. △ Less

Submitted 9 May, 2026; originally announced May 2026.

Comments: 40 pages, 45 figures

A Single-Loop Penalty-based Algorithm for Stochastic Minimax Optimization with Nonlinear Coupled Constraints

Authors: Qichao Cao, Shangzhi Zeng, Jin Zhang, Yuxuan Zhou

Abstract: We study stochastic nonconvex-concave minimax optimization with nonlinear coupled constraints that are convex in the maximization variable. To address the nonsmoothness arising from such constraints, we develop a penalty-based smooth approximation that combines quadratic penalization of the coupled constraints with quadratic regularization of the inner maximization problem. Based on this approxima… ▽ More We study stochastic nonconvex-concave minimax optimization with nonlinear coupled constraints that are convex in the maximization variable. To address the nonsmoothness arising from such constraints, we develop a penalty-based smooth approximation that combines quadratic penalization of the coupled constraints with quadratic regularization of the inner maximization problem. Based on this approximation, we propose SPACO, a single-loop stochastic gradient algorithm that tracks the inner maximizer by one stochastic ascent step, updates the outer variable using an inexact stochastic descent direction, and adaptively updates the penalty and regularization parameters over the iterations. For the penalty-based smooth approximation, we establish convergence guarantees from both minimizer and stationarity perspectives. In particular, we introduce enhanced KKT conditions and show that stationary points of the smooth approximations can converge to points satisfying these conditions. An example illustrates that the enhanced KKT conditions can help exclude KKT points that are not local minimizers. For SPACO, we prove non-asymptotic complexity bounds for stationarity and feasibility, as well as asymptotic subsequential convergence to enhanced KKT points. Numerical experiments on synthetic examples, fairness-aware classification, and constrained generative adversarial network training demonstrate the effectiveness of the proposed method. △ Less

Submitted 8 June, 2026; v1 submitted 2 May, 2026; originally announced May 2026.

Exponentially improved quantum simulation of scalar QFT

Authors: Qing-Hong Cao, Ying-Ying Li, Xiaohui Liu, Liang-Qi Zhang, Ke Zhao

Abstract: Quantum simulations of scalar quantum field theories (QFT) provide important benchmarks for demonstrating quantum advantage. We revisit digitization in the occupation basis, which is typically hindered by unfavorable circuit depth scaling. We present an approach that achieves exponential reductions in circuit depth and significantly mitigates Trotter errors by diagonalizing field operators prior t… ▽ More Quantum simulations of scalar quantum field theories (QFT) provide important benchmarks for demonstrating quantum advantage. We revisit digitization in the occupation basis, which is typically hindered by unfavorable circuit depth scaling. We present an approach that achieves exponential reductions in circuit depth and significantly mitigates Trotter errors by diagonalizing field operators prior to their decomposition into Pauli strings. Focusing on a scalar QFT in 2+1 dimensions, we show that this method substantially reduces circuit depth and CNOT gate counts for time evolution. Using the Lorentzian energy-energy correlator as a benchmark observable, we find parameter regimes in which occupation-basis digitization converges more rapidly with respect to local truncation than the amplitude-basis approach of Jordan, Lee, and Preskill. These results provide both algorithmic advances and phenomenological benchmarks for studies of light-ray observables on near-term quantum devices. △ Less

Submitted 28 April, 2026; originally announced April 2026.

Comments: 11 pages,5 figures,6 tables

The First Challenge on Remote Sensing Infrared Image Super-Resolution at NTIRE 2026: Benchmark Results and Method Overview

Authors: Kai Liu, Haoyang Yue, Zeli Lin, Zheng Chen, Jingkai Wang, Jue Gong, Jiatong Li, Xianglong Yan, Libo Zhu, Jianze Li, Ziqing Zhang, Zihan Zhou, Xiaoyang Liu, Radu Timofte, Yulun Zhang, Junye Chen, Zhenming Yan, Yucong Hong, Ruize Han, Song Wang, Li Pang, Heng Zhao, Xinqiao Wu, Deyu Meng, Xiangyong Cao , et al. (43 additional authors not shown)

Abstract: This paper presents the NTIRE 2026 Remote Sensing Infrared Image Super-Resolution (x4) Challenge, one of the associated challenges of NTIRE 2026. The challenge aims to recover high-resolution (HR) infrared images from low-resolution (LR) inputs generated through bicubic downsampling with a x4 scaling factor. The objective is to develop effective models or solutions that achieve state-of-the-art pe… ▽ More This paper presents the NTIRE 2026 Remote Sensing Infrared Image Super-Resolution (x4) Challenge, one of the associated challenges of NTIRE 2026. The challenge aims to recover high-resolution (HR) infrared images from low-resolution (LR) inputs generated through bicubic downsampling with a x4 scaling factor. The objective is to develop effective models or solutions that achieve state-of-the-art performance for infrared image SR in remote sensing scenarios. To reflect the characteristics of infrared data and practical application needs, the challenge adopts a single-track setting. A total of 115 participants registered for the competition, with 13 teams submitting valid entries. This report summarizes the challenge design, dataset, evaluation protocol, main results, and the representative methods of each team. The challenge serves as a benchmark to advance research in infrared image super-resolution and promote the development of effective solutions for real-world remote sensing applications. △ Less

Submitted 23 April, 2026; originally announced April 2026.

Comments: Github Repo: https://github.com/Kai-Liu001/NTIRE2026_infraredSR

The nEXO Radioassay Program

Authors: R. MacLellan, P. Acharya, B. Aharmim, S. Alcantar Anguiano, A. Anker, I. J. Arnquist, D. Auty, T. Bhatta, D. Chernyak, J. S. Choe, B. Cleveland, J. Daughhetee, A. Der Mesrobian-Kabakian, Y. Y. Ding, M. L. di Vacri, J. Farine, A. D. French, O. Gileva, R. Gornea, K. Harouaka, K. P. Hobbs, E. W. Hoppe, L. K. S. Horkley, M. Hughes, L. Kieser , et al. (126 additional authors not shown)

Abstract: Material radioactivity compilations, such as the one presented here, are important enablers of science. They are useful for the selection of radiopure materials used in the design and construction of low-energy rare-event search experiments. They allow researchers developing such experiments to save time on material studies and avoid costly duplication of effort. The data presented here were gener… ▽ More Material radioactivity compilations, such as the one presented here, are important enablers of science. They are useful for the selection of radiopure materials used in the design and construction of low-energy rare-event search experiments. They allow researchers developing such experiments to save time on material studies and avoid costly duplication of effort. The data presented here were generated in support of the planned nEXO double-beta decay search. This work contains among the most restrictive constraints on the natural radioactivity content of materials of general interest to the low-radioactivity community, found in any tabulation of this kind. In this study, various techniques were employed; they are described here. △ Less

Submitted 16 April, 2026; originally announced April 2026.

The Fourth Challenge on Image Super-Resolution ($\times$4) at NTIRE 2026: Benchmark Results and Method Overview

Authors: Zheng Chen, Kai Liu, Jingkai Wang, Xianglong Yan, Jianze Li, Ziqing Zhang, Jue Gong, Jiatong Li, Lei Sun, Xiaoyang Liu, Radu Timofte, Yulun Zhang, Jihye Park, Yoonjin Im, Hyungju Chun, Hyunhee Park, MinKyu Park, Zheng Xie, Xiangyu Kong, Weijun Yuan, Zhan Li, Qiurong Song, Luen Zhu, Fengkai Zhang, Xinzhe Zhu , et al. (128 additional authors not shown)

Abstract: This paper presents the NTIRE 2026 image super-resolution ($\times$4) challenge, one of the associated competitions of the NTIRE 2026 Workshop at CVPR 2026. The challenge aims to reconstruct high-resolution (HR) images from low-resolution (LR) inputs generated through bicubic downsampling with a $\times$4 scaling factor. The objective is to develop effective super-resolution solutions and analyze… ▽ More This paper presents the NTIRE 2026 image super-resolution ($\times$4) challenge, one of the associated competitions of the NTIRE 2026 Workshop at CVPR 2026. The challenge aims to reconstruct high-resolution (HR) images from low-resolution (LR) inputs generated through bicubic downsampling with a $\times$4 scaling factor. The objective is to develop effective super-resolution solutions and analyze recent advances in the field. To reflect the evolving objectives of image super-resolution, the challenge includes two tracks: (1) a restoration track, which emphasizes pixel-wise fidelity and ranks submissions based on PSNR; and (2) a perceptual track, which focuses on visual realism and evaluates results using a perceptual score. A total of 194 participants registered for the challenge, with 31 teams submitting valid entries. This report summarizes the challenge design, datasets, evaluation protocol, main results, and methods of participating teams. The challenge provides a unified benchmark and offers insights into current progress and future directions in image super-resolution. △ Less

Submitted 15 April, 2026; originally announced April 2026.

Comments: NTIRE 2026 webpage: https://cvlai.net/ntire/2026. Code: https://github.com/zhengchen1999/NTIRE2026_ImageSR_x4

AIBuildAI: An AI Agent for Automatically Building AI Models

Authors: Ruiyi Zhang, Peijia Qin, Qi Cao, Li Zhang, Pengtao Xie

Abstract: AI models underpin modern intelligent systems, driving advances across science, medicine, finance, and technology. Yet developing high-performing AI models remains a labor-intensive process that requires expert practitioners to iteratively design architectures, engineer representations, implement training pipelines and refine approaches through empirical evaluation. Existing AutoML methods partial… ▽ More AI models underpin modern intelligent systems, driving advances across science, medicine, finance, and technology. Yet developing high-performing AI models remains a labor-intensive process that requires expert practitioners to iteratively design architectures, engineer representations, implement training pipelines and refine approaches through empirical evaluation. Existing AutoML methods partially alleviate this burden but remain limited to narrow aspects such as hyperparameter optimization and model selection within predefined search spaces, leaving the full development lifecycle largely dependent on human expertise. To address this gap, we introduce AIBuildAI, an AI agent that automatically builds AI models from a task description and training data. AIBuildAI adopts a hierarchical agent architecture in which a manager agent coordinates three specialized sub-agents: a designer for modeling strategy, a coder for implementation and debugging, and a tuner for training and performance optimization. Each sub-agent is itself a large language model (LLM) based agent capable of multi-step reasoning and tool use, enabling end-to-end automation of the AI model development process that goes beyond the scope of existing AutoML approaches. We evaluate AIBuildAI on MLE-Bench, a benchmark of realistic Kaggle-style AI development tasks spanning visual, textual, time-series and tabular modalities. AIBuildAI ranks first on MLE-Bench with a medal rate of 63.1%, outperforming all existing baseline methods and matching the capability of highly experienced AI engineers. These results demonstrate that hierarchical agent systems can automate the full AI model development process from task specification to deployable model, suggesting a pathway toward broadly accessible AI development with minimal human intervention. △ Less

Submitted 15 April, 2026; originally announced April 2026.

Feed-Forward 3D Scene Modeling: A Problem-Driven Perspective

Authors: Weijie Wang, Qihang Cao, Sensen Gao, Donny Y. Chen, Haofei Xu, Wenjing Bian, Songyou Peng, Tat-Jen Cham, Chuanxia Zheng, Andreas Geiger, Jianfei Cai, Jia-Wang Bian, Bohan Zhuang

Abstract: Reconstructing 3D representations from 2D inputs is a fundamental task in computer vision and graphics, serving as a cornerstone for understanding and interacting with the physical world. While traditional methods achieve high fidelity, they are limited by slow per-scene optimization or category-specific training, which hinders their practical deployment and scalability. Hence, generalizable feed-… ▽ More Reconstructing 3D representations from 2D inputs is a fundamental task in computer vision and graphics, serving as a cornerstone for understanding and interacting with the physical world. While traditional methods achieve high fidelity, they are limited by slow per-scene optimization or category-specific training, which hinders their practical deployment and scalability. Hence, generalizable feed-forward 3D reconstruction has witnessed rapid development in recent years. By learning a model that maps images directly to 3D representations in a single forward pass, these methods enable efficient reconstruction and robust cross-scene generalization. Our survey is motivated by a critical observation: despite the diverse geometric output representations, ranging from implicit fields to explicit primitives, existing feed-forward approaches share similar high-level architectural patterns, such as image feature extraction backbones, multi-view information fusion mechanisms, and geometry-aware design principles. Consequently, we abstract away from these representation differences and instead focus on model design, proposing a novel taxonomy centered on model design strategies that are agnostic to the output format. Our proposed taxonomy organizes the research directions into five key problems that drive recent research development: feature enhancement, geometry awareness, model efficiency, augmentation strategies and temporal-aware models. To support this taxonomy with empirical grounding and standardized evaluation, we further comprehensively review related benchmarks and datasets, and extensively discuss and categorize real-world applications based on feed-forward 3D models. Finally, we outline future directions to address open challenges such as scalability, evaluation standards, and world modeling. △ Less

Submitted 15 April, 2026; originally announced April 2026.

Comments: 67 pages, 395 references. Project page: https://ff3d-survey.github.io. Code: https://github.com/ziplab/Awesome-Feed-Forward-3D. This work has been submitted to Springer for possible publication

Analogical Reasoning as a Doctor: A Foundation Model for Gastrointestinal Endoscopy Diagnosis

Authors: Peixi Peng, Housheng Xie, Yanling Wei, Guangcong Ruan, Xiaoyang Zou, Qian Cao, Yongjian Nian, Guoyan Zheng

Abstract: Gastrointestinal diseases impose a growing global health burden, and endoscopy is a primary tool for early diagnosis. However, routine endoscopic image interpretation still suffers from missed lesions and limited efficiency. Although AI-assisted diagnosis has shown promise, existing models often lack generalizability, adaptability, robustness, and scalability because of limited medical data, domai… ▽ More Gastrointestinal diseases impose a growing global health burden, and endoscopy is a primary tool for early diagnosis. However, routine endoscopic image interpretation still suffers from missed lesions and limited efficiency. Although AI-assisted diagnosis has shown promise, existing models often lack generalizability, adaptability, robustness, and scalability because of limited medical data, domain shift, and heterogeneous annotations. To address these challenges, we develop RATNet, a foundation model for gastrointestinal endoscopy imaging based on analogical reasoning. RATNet acquires and transfers knowledge from heterogeneous expert annotations across five gastrointestinal endoscopy datasets through a cyclic pre-training strategy. Its architecture consists of an encoder, a relevance-knowledge acquisition and transfer (RAT) module, a projector, and a multi-task head, and supports fine-tuning, linear probing, and zero-shot transfer. Evaluations show that RATNet outperforms existing foundation models, including GastroNet and GastroVision, across six scenarios: diagnosis of common gastrointestinal diseases, few-shot learning for rare diseases, zero-shot transfer to new medical sites, robustness under long-tailed disease distributions, adaptation to novel diseases, and privacy-preserving deployment via federated learning. Its advantage comes from an analogical reasoning mechanism that matches image-derived posterior knowledge to a learned prior knowledge base and transfers relative knowledge to guide diagnosis, improving generalization and resistance to bias. RATNet is open and cost-effective, supports automatic integration of heterogeneous annotations without manual label unification, and reduces data acquisition costs, making it a practical foundation for intelligent gastrointestinal diagnosis, especially in resource-limited settings. △ Less

Submitted 7 April, 2026; originally announced April 2026.

Instruction-Tuned LLMs for Parsing and Mining Unstructured Logs on Leadership HPC Systems

Authors: Ahmad Maroof Karimi, Jong Youl Choi, Charles Qing Cao, Awais Khan

Abstract: Leadership-class HPC systems generate massive volumes of heterogeneous, largely unstructured system logs. Because these logs originate from diverse software, hardware, and runtime layers, they exhibit inconsistent formats, making structure extraction and pattern discovery extremely challenging. Therefore, robust log parsing and mining is critical to transform this raw telemetry into actionable ins… ▽ More Leadership-class HPC systems generate massive volumes of heterogeneous, largely unstructured system logs. Because these logs originate from diverse software, hardware, and runtime layers, they exhibit inconsistent formats, making structure extraction and pattern discovery extremely challenging. Therefore, robust log parsing and mining is critical to transform this raw telemetry into actionable insights that reveal operational patterns, diagnose anomalies, and enable reliable, efficient, and scalable system analysis. Recent advances in large language models (LLMs) offer a promising new direction for automated log understanding in leadership-class HPC environments. To capitalize on this opportunity, we present a domain-adapted, instruction-following, LLM-driven framework that leverages chain-of-thought (CoT) reasoning to parse and structure HPC logs with high fidelity. Our approach combines domain-specific log-template data with instruction-tuned examples to fine-tune an 8B-parameter LLaMA model tailored for HPC log analysis. We develop a hybrid fine-tuning methodology that adapts a general-purpose LLM to domain-specific log data, enabling privacy-preserving, locally deployable, fast, and energy-efficient log-mining approach. We conduct experiments on a diverse set of log datasets from the LogHub repository. The evaluation confirms that our approach achieves parsing accuracy on par with significantly larger models, such as LLaMA 70B and Anthropic's Claude. We further validate the practical utility of our fine-tuned LLM model by parsing over 600 million production logs from the Frontier supercomputer over a four-week window, uncovering critical patterns in temporal dynamics, node-level anomalies, and workload-error log correlations. △ Less

Submitted 6 April, 2026; originally announced April 2026.

NTIRE 2026 3D Restoration and Reconstruction in Real-world Adverse Conditions: RealX3D Challenge Results

Authors: Shuhong Liu, Chenyu Bao, Ziteng Cui, Xuangeng Chu, Bin Ren, Lin Gu, Xiang Chen, Mingrui Li, Long Ma, Marcos V. Conde, Radu Timofte, Yun Liu, Ryo Umagami, Tomohiro Hashimoto, Zijian Hu, Yuan Gan, Tianhan Xu, Yusuke Kurose, Tatsuya Harada, Junwei Yuan, Gengjia Chang, Xining Ge, Mache You, Qida Cao, Zeliang Li , et al. (81 additional authors not shown)

Abstract: This paper presents a comprehensive review of the NTIRE 2026 3D Restoration and Reconstruction (3DRR) Challenge, detailing the proposed methods and results. The challenge seeks to identify robust reconstruction pipelines that are robust under real-world adverse conditions, specifically extreme low-light and smoke-degraded environments, as captured by our RealX3D benchmark. A total of 279 participa… ▽ More This paper presents a comprehensive review of the NTIRE 2026 3D Restoration and Reconstruction (3DRR) Challenge, detailing the proposed methods and results. The challenge seeks to identify robust reconstruction pipelines that are robust under real-world adverse conditions, specifically extreme low-light and smoke-degraded environments, as captured by our RealX3D benchmark. A total of 279 participants registered for the competition, of whom 33 teams submitted valid results. We thoroughly evaluate the submitted approaches against state-of-the-art baselines, revealing significant progress in 3D reconstruction under adverse conditions. Our analysis highlights shared design principles among top-performing methods and provides insights into effective strategies for handling 3D scene degradation. △ Less

Submitted 29 April, 2026; v1 submitted 5 April, 2026; originally announced April 2026.

JoyAI-LLM Flash: Advancing Mid-Scale LLMs with Token Efficiency

Authors: Aichen Cai, Anmeng Zhang, Anyu Li, Bo Zhang, Bohua Cai, Chang Li, Changjian Jiang, Changkai Lu, Chao Xue, Chaocai Liang, Cheng Zhang, Dongkai Liu, Fei Wang, Guoqiang Huang, Haijian Ke, Han Lin, Hao Wang, Ji Miao, Jiacheng Zhang, Jialong Shi, Jifeng Zhu, Jingjing Qian, Junhui Luo, Junwu Xiong, Lam So , et al. (44 additional authors not shown)

Abstract: We introduce JoyAI-LLM Flash, an efficient Mixture-of-Experts (MoE) language model designed to redefine the trade-off between strong performance and token efficiency in the sub-50B parameter regime. JoyAI-LLM Flash is pretrained on a massive corpus of 20 trillion tokens and further optimized through a rigorous post-training pipeline, including supervised fine-tuning (SFT), Direct Preference Optimi… ▽ More We introduce JoyAI-LLM Flash, an efficient Mixture-of-Experts (MoE) language model designed to redefine the trade-off between strong performance and token efficiency in the sub-50B parameter regime. JoyAI-LLM Flash is pretrained on a massive corpus of 20 trillion tokens and further optimized through a rigorous post-training pipeline, including supervised fine-tuning (SFT), Direct Preference Optimization (DPO), and large-scale reinforcement learning (RL) across diverse environments. To improve token efficiency, JoyAI-LLM Flash strategically balances \emph{thinking} and \emph{non-thinking} cognitive modes and introduces FiberPO, a novel RL algorithm inspired by fibration theory that decomposes trust-region maintenance into global and local components, providing unified multi-scale stability control for LLM policy optimization. To enhance architectural sparsity, the model comprises 48B total parameters while activating only 2.7B parameters per forward pass, achieving a substantially higher sparsity ratio than contemporary industry leading models of comparable scale. To further improve inference throughput, we adopt a joint training-inference co-design that incorporates dense Multi-Token Prediction (MTP) and Quantization-Aware Training (QAT). We release the checkpoints for both JoyAI-LLM-48B-A3B Base and its post-trained variants on Hugging Face to support the open-source community. △ Less

Submitted 8 April, 2026; v1 submitted 3 April, 2026; originally announced April 2026.

Comments: Xiaodong He is the corresponding author

GenSmoke-GS: A Multi-Stage Method for Novel View Synthesis from Smoke-Degraded Images Using a Generative Model

Authors: Qida Cao, Xinyuan Hu, Changyue Shi, Jiajun Ding, Zhou Yu, Jun Yu

Abstract: This paper describes our method for Track 2 of the NTIRE 2026 3D Restoration and Reconstruction (3DRR) Challenge on smoke-degraded images. In this task, smoke reduces image visibility and weakens the cross-view consistency required by scene optimization and rendering. We address this problem with a multi-stage pipeline consisting of image restoration, dehazing, MLLM-based enhancement, 3DGS-MCMC op… ▽ More This paper describes our method for Track 2 of the NTIRE 2026 3D Restoration and Reconstruction (3DRR) Challenge on smoke-degraded images. In this task, smoke reduces image visibility and weakens the cross-view consistency required by scene optimization and rendering. We address this problem with a multi-stage pipeline consisting of image restoration, dehazing, MLLM-based enhancement, 3DGS-MCMC optimization, and averaging over repeated runs. The main purpose of the pipeline is to improve visibility before rendering while limiting scene-content changes across input views. Experimental results on the challenge benchmark show improved quantitative performance and better visual quality than the provided baselines. The code is available at https://github.com/plbbl/GenSmoke-GS. Our method achieved a ranking of 1 out of 14 participants in Track 2 of the NTIRE 3DRR Challenge, as reported on the official competition website: https://www.codabench.org/competitions/13993/#/results-tab. △ Less

Submitted 6 April, 2026; v1 submitted 3 April, 2026; originally announced April 2026.

PRBench: End-to-end Paper Reproduction in Physics Research

Authors: Shi Qiu, Junyi Deng, Yiwei Deng, Haoran Dong, Jieyu Fu, Mao Li, Zeyu Li, Zhaolong Zhang, Huiwen Zheng, Leidong Bao, Anqi Lv, Zihan Mo, Yadi Niu, Yiyang Peng, Yu Tian, Yili Wang, Ziyu Wang, Zi-Yu Wang, Jiashen Wei, Liuheng Wu, Aoran Xue, Leyi Yang, Guanglu Yuan, Xiarui Zhan, Jingjun Zhang , et al. (26 additional authors not shown)

Abstract: AI agents powered by large language models exhibit strong reasoning and problem-solving capabilities, enabling them to assist scientific research tasks such as formula derivation and code generation. However, whether these agents can reliably perform end-to-end reproduction from real scientific papers remains an open question. We introduce PRBench, a benchmark of 30 expert-curated tasks spanning 1… ▽ More AI agents powered by large language models exhibit strong reasoning and problem-solving capabilities, enabling them to assist scientific research tasks such as formula derivation and code generation. However, whether these agents can reliably perform end-to-end reproduction from real scientific papers remains an open question. We introduce PRBench, a benchmark of 30 expert-curated tasks spanning 11 subfields of physics. Each task requires an agent to comprehend the methodology of a published paper, implement the corresponding algorithms from scratch, and produce quantitative results matching the original publication. Agents are provided only with the task instruction and paper content, and operate in a sandboxed execution environment. All tasks are contributed by domain experts from over 20 research groups at the School of Physics, Peking University, each grounded in a real published paper and validated through end-to-end reproduction with verified ground-truth results and detailed scoring rubrics. Using an agentified assessment pipeline, we evaluate a set of coding agents on PRBench and analyze their capabilities across key dimensions of scientific reasoning and execution. The best-performing agent, OpenAI Codex powered by GPT-5.3-Codex, achieves a mean overall score of 34%. All agents exhibit a zero end-to-end callback success rate, with particularly poor performance in data accuracy and code correctness. We further identify systematic failure modes, including errors in formula implementation, inability to debug numerical simulations, and fabrication of output data. Overall, PRBench provides a rigorous benchmark for evaluating progress toward autonomous scientific research. △ Less

Submitted 29 March, 2026; originally announced March 2026.

Comments: 17 pages, 3 figures

Report number: RISE-AGI-2026-002

Statistical modeling of breast cancer radiomic features and hazard using image registration-aided longitudinal CT data

Authors: Subrata Mukherjee, Qian Cao, Thibaud Coroller, Ravi K. Samala, Nicholas Petrick, Berkman Sahiner

Abstract: Patients with metastatic breast cancer (mBC) undergo repeated computed tomography (CT) imaging during treatment to monitor disease progression. Accurate longitudinal tracking of individual lesions across scans from multiple radiologists is essential for reliable radiomic analysis and clinical decision-making. We conducted a retrospective study using serial chest CT scans from the Phase III MONALEE… ▽ More Patients with metastatic breast cancer (mBC) undergo repeated computed tomography (CT) imaging during treatment to monitor disease progression. Accurate longitudinal tracking of individual lesions across scans from multiple radiologists is essential for reliable radiomic analysis and clinical decision-making. We conducted a retrospective study using serial chest CT scans from the Phase III MONALEESA-3 and MONALEESA-7 trials and developed statistical models for multi-source data integration and survival analysis. First, we introduced a Registration-based Automated Matching and Correspondence (RAMAC) algorithm to establish lesion correspondence across annotations from different radiologists and imaging time points using the Hungarian algorithm. Second, using the RAMAC-processed dataset, we developed interpretable radiomic survival models for progression-free survival prediction by combining baseline radiomic features, post-treatment changes at Weeks 8, 16, and 24, and demographic variables. To address the high dimensionality of longitudinal radiomic data, feature reduction was performed using an L1-penalized additive Cox proportional hazards model and best subset selection followed by Cox modeling. Model performance was evaluated using the concordance index (C-index). Incorporating additional imaging time points improved predictive performance, increasing the mean C-index from 0.58 at baseline to 0.64. Joint modeling further showed significant associations between longitudinal radiomic features and survival outcomes over time. △ Less

Submitted 27 March, 2026; originally announced March 2026.

Comments: 18 pages, 8 figure, 2 tables. arXiv admin note: substantial text overlap with arXiv:2501.06814