Showing 1–50 of 109 results for author: Chi, J

HarnessEval-W: Agentifying the Evaluation of Visual Worlds

Authors: Weiliang Chen, Haowen Sun, Jun Gao, Jiawei Chi, Hanyang Wang, Qiyu Dai, Yihao Li, Hao Li, Jingnan Gao, Yi-Hsin Hung, Xingzhuo Guo, Shangchen Miao, Zhiyuan Shi, Xiang Li, Fengrui Tian, Weihua Du, Ziqi Huang, Shenyuan Gao, Siqiao Huang, Mingyu Liu, Yifei Li, Shizun Wang, Xi Wang, Tianqi Zhang, Xue Luo , et al. (18 additional authors not shown)

Abstract: A benchmark should deliver more than a scalar score: what makes an evaluation trustworthy is the reasoning that justifies the score. This is especially critical for world models, where judging a rollout requires understanding whether physics, causality, and world state evolve correctly. Humans spot such violations naturally, yet no existing benchmark automates this capability: metrics are computed… ▽ More A benchmark should deliver more than a scalar score: what makes an evaluation trustworthy is the reasoning that justifies the score. This is especially critical for world models, where judging a rollout requires understanding whether physics, causality, and world state evolve correctly. Humans spot such violations naturally, yet no existing benchmark automates this capability: metrics are computed brute-force, leaving no reasoning chain that can be examined or verified. We introduce HarnessEval-W, an agentified evaluation pipeline that brings the harness paradigm from the LLM ecosystem to world model benchmarking. Rather than applying a fixed rubric, HarnessEval-W interprets the context of each evaluation case, decomposes the evaluation question into measurable subproblems, and spawns specialized sub-agents, each equipped with tailored context and diagnostic tools to reason over its own subproblem. The parent agent then validates the gathered evidence and summarizes it into the final verdict. This hierarchical workflow turns every evaluation into a transparent evidence tree whose complete reasoning chain justifies the result. We apply HarnessEval-W to 18 representative world models over 330 evaluation cases. Its judgments closely align with human preferences while providing verifiable, fine-grained diagnoses of every generated rollout. We open-source the full pipeline as a live benchmark and invite the broad community to contribute to grow new skills and evaluation cases as world models evolve. △ Less

Submitted 17 August, 2026; originally announced August 2026.

Comments: Project Page: https://mirros-lab.github.io/HarnessEval-W

Multidimensional Design of Metal-Nitrogen-Carbon Electrocatalysts for Direct Propylene Epoxidation

Authors: Songbo Ye, Qingyuan Han, Jingwen Chi, Yuan Huang, Heng Liu, Di Zhang, Hitoshi Shiku, Li Wei, Hao Li

Abstract: Propylene oxide is a major industrial chemical whose production currently relies on hazardous chlorine- or peroxide-based oxidants. Direct electrochemical epoxidation using water as the oxygen source offers a sustainable alternative, but controlling oxygen-atom transfer against the competing oxygen evolution reaction remains a fundamental challenge. Here, we show that propylene epoxidation selecti… ▽ More Propylene oxide is a major industrial chemical whose production currently relies on hazardous chlorine- or peroxide-based oxidants. Direct electrochemical epoxidation using water as the oxygen source offers a sustainable alternative, but controlling oxygen-atom transfer against the competing oxygen evolution reaction remains a fundamental challenge. Here, we show that propylene epoxidation selectivity cannot be described by oxygen binding energy alone, but is jointly governed by oxygen adsorption, the potential of zero charge, and applied potential. By combining theoretical calculations with pH-field-coupled microkinetic modeling across 41 metal-nitrogen-carbon single-atom catalysts, we first identified an optimal oxygen-binding window and Co as the most favorable metal center. We then found that peripheral substituents can tune the PZC while largely preserving the optimal oxygen adsorption energetics, thereby providing an independent design dimension to further optimize the already favorable Co active site. This sequential, multidimensional design strategy identified CoPc-NH2-CNT as the optimal catalyst, delivering a record PO Faradaic efficiency of 70-80 percent for direct propylene epoxidation in aqueous electrolyte under ambient conditions. These results establish interfacial electrostatics as an independently tunable design dimension for controlling selective oxygen-atom transfer in electrocatalysis. △ Less

Submitted 14 August, 2026; originally announced August 2026.

Comments: 19 pages

A Multi-Layer System for Ultra-High-Resolution Static 360-Degree Telepresence

Authors: Jiapeng Chi, Gerd Bruder, Carsten Neumann, Carolina Cruz-Neira, Dirk Reiners

Abstract: 360-degree video telepresence offers strong immersive potential but remains constrained by the limited resolution of current capture and display hardware. Many telepresence installations feature fixed viewpoints and largely static scenes, yet optimization strategies tailored to such setups have received limited attention. We present a multi-layer, ultra-high-resolution system for static 360-degree… ▽ More 360-degree video telepresence offers strong immersive potential but remains constrained by the limited resolution of current capture and display hardware. Many telepresence installations feature fixed viewpoints and largely static scenes, yet optimization strategies tailored to such setups have received limited attention. We present a multi-layer, ultra-high-resolution system for static 360-degree telepresence that combines an 8K panoramic camera with a rotatable 4K pan-tilt-zoom (PTZ) camera. Our approach builds a three-layer representation: (1) a tile-based ultra-high-resolution panoramic background, generated by offline stitching high-detail 4K PTZ scans onto the base 8K panorama to achieve effective resolution beyond native capture, and represented as a set of spatial tiles; (2) a dynamic update layer that composites foreground motions from the 8K stream via real-time high-resolution background matting; and (3) a region-of-interest 4K layer that streams a real-time PTZ view of the selected region and additionally updates the corresponding background tiles over time. We evaluate the proposed system through comparisons with representative video super-resolution approaches and a user study assessing perceived detail and immersive experience. Our results indicate that tile-based background refinement, together with user-guided updates, provides a practical way to balance panoramic fidelity and interactivity in static 360-degree telepresence. △ Less

Submitted 5 August, 2026; originally announced August 2026.

Comments: 11 pages, 11 figures. Accepted to IEEE ISMAR 2026, to appear in IEEE Transactions on Visualization and Computer Graphics (TVCG)

AgenticASR: Refining Speech Recognition in Real-World Scenarios via an Agentic Approach

Authors: Zixuan Jiang, Binghao Qiang, Jiaying Chi, Yanqiao Zhu, Kai Yu, Xie Chen

Abstract: Automatic speech recognition (ASR) has achieved substantial gains in transcription accuracy, yet verbatim transcription does not necessarily produce readily usable text. It retains fillers, repetitions, false starts, and self-corrections that increase reading effort, obscure the speaker's final intent, and propagate unresolved or abandoned content to downstream tasks. Existing spoken-to-written me… ▽ More Automatic speech recognition (ASR) has achieved substantial gains in transcription accuracy, yet verbatim transcription does not necessarily produce readily usable text. It retains fillers, repetitions, false starts, and self-corrections that increase reading effort, obscure the speaker's final intent, and propagate unresolved or abandoned content to downstream tasks. Existing spoken-to-written methods process completed audio or transcripts but cannot revise emitted text when later speech changes how preceding content should be interpreted. We therefore formulate Agentic Speech Recognition (AgenticSR), an audio-to-clean-text task that removes disfluencies, resolves self-corrections, and normalizes written form while preserving the speaker's final intent. AgenticASR implements this task through an ASR--Refiner architecture that repeatedly transforms a bounded active context and replaces its corresponding output span as audio arrives. This enables continual emission and revision over streams of arbitrary duration. We also introduce AASR-Bench, a bilingual benchmark with fine-grained atomic rubrics. Across multiple ASR front ends, AgenticASR attains the highest AASR-Bench scores among evaluated systems. A human--AI agreement study shows that rubric-based judgments align with independent expert assessments. Ablations characterize Refiner capacity, context length, and the quality--latency trade-off between online and offline inference. Together, these results establish AgenticASR as a practical framework for intent-preserving clean transcription during ongoing speech. Code, AASR-Bench, and a demo will be released at https://github.com/AnXMuy/AgenticASR. △ Less

Submitted 30 July, 2026; originally announced July 2026.

Comments: 15 pages, 3 figures, 14 tables

Food Image Segmentation with LLM-Derived Ingredient Labels and Multimodal Fusion

Authors: Jui-Feng Chi, Wei-Ta Chu, Sheng-Long Lin

Abstract: Food image segmentation plays a vital role in health-related applications such as nutrition tracking and personalized health monitoring. However, existing models often underperform on visually similar ingredients and rare food categories. To address this issue, we propose two plug-and-play multimodal modules that enhance the segmentation performance by leveraging ingredient labels inferred from fo… ▽ More Food image segmentation plays a vital role in health-related applications such as nutrition tracking and personalized health monitoring. However, existing models often underperform on visually similar ingredients and rare food categories. To address this issue, we propose two plug-and-play multimodal modules that enhance the segmentation performance by leveraging ingredient labels inferred from food images using large language models (LLMs). The first module, called LIM-F (Language Injection Module for Features), is designed to pair with any image encoder that produces multi-layer outputs (e.g., Swin Transformer), while the second module, LIM-Q (Language Injection Module for Queries), targets Mask2Former-style Transformer-based decoders. Both modules enable training without the need for pre-aligning images with text by directly injecting semantic ingredient information into the visual analysis pipeline. On the FoodSeg103 benchmark, the proposed method achieves state-of-the-art performance. Specifically, integrating LIM-Q into the Mask2Former decoder with a Swin-L image encoder yields a mean Intersection over Union (mIoU) of 55.0. LIM-F also demonstrates strong generalization and competitive performance, reaching an mIoU of 54.4 under the same model (Swin-L+Mask2Former). Furthermore, its applicability extends beyond Transformer-based decoders, as evidenced by an improvement from 47.7 to 49.8 mIoU when integrated into a CNN-based architecture. Notably, the improved segmentation accuracy is achieved with only a moderate (at most 3.8 GB) increase in the GPU memory consumption during training. Thus, the proposed approach offers a practical and scalable solution for fine-grained food understanding. △ Less

Submitted 28 July, 2026; originally announced July 2026.

Fine-Grained Food Image Understanding via Target-Aware Data Alignment

Authors: Jui-Feng Chi, Wei-Lun Chu, Bruce Coburn, Jinge Ma, Fengqing Zhu

Abstract: Fine-grained food visual--semantic understanding requires models to capture subtle distinctions across ingredients, cooking methods, doneness, color, texture, and plate composition. Although CLIP-style vision-language models provide a natural framework for this task, their effectiveness is limited when training relies on heterogeneous web-collected image--text pairs. Such data often exhibit a web-… ▽ More Fine-grained food visual--semantic understanding requires models to capture subtle distinctions across ingredients, cooking methods, doneness, color, texture, and plate composition. Although CLIP-style vision-language models provide a natural framework for this task, their effectiveness is limited when training relies on heterogeneous web-collected image--text pairs. Such data often exhibit a web-to-target domain gap and cross-modal misalignment, where images differ from the target distribution and captions are noisy, multilingual, or weakly grounded in visual content. We propose a data-centric multimodal alignment method for fine-grained food description and recognition. Our method first performs target-aware data selection to identify visually relevant training subsets, then applies VLM-based caption refinement to generate visually grounded, target-style descriptions. Using these curated image--caption pairs, we train complementary CLIP-style retrieval experts and further combine their decisions through a hierarchical VLM-assisted multi-expert decision-level fusion strategy that invokes the VLM only when experts disagree. Experiments show that our data refinement strategy significantly improves retrieval performance over naive web supervision, with VLM-based caption refinement alone yielding an average performance gain of approximately 19%. Our full method also achieves more than twice the retrieval score of pure VLM-based retrieval while remaining substantially more efficient. △ Less

Submitted 28 July, 2026; originally announced July 2026.

FinSAgent: Corpus-Aligned Multi-Agent RAG Framework for Evidence-Grounded SEC Filing Question Answering

Authors: Jijun Chi, Zhenghan Tai, Hanwei Wu, Tung Sum Thomas Kwok, Hailin He, Zixing Liao, Bohuai Xiao, Chaolong Jiang, Jianliang Lei, Jerry Huang, Peng Lu, Muzhi Li, Liheng Ma, Yihong Wu, Sicheng Lyu, Jingrui Tian, Yihan Li, Yanzhang Ma, Sizhe Guan, Dingtao Hu, Yufei Cui, Ling Zhou, Lei Ding, Xinyu Wang

Abstract: Financial question answering over U.S. Securities and Exchange Commission (SEC) filings requires retrieving and synthesizing heterogeneous evidence dispersed across long, standardized, and highly redundant disclosures. Existing retrieval-augmented and multi-agent systems typically derive retrieval queries directly from the user's question and rank candidates by semantic similarity. Together, these… ▽ More Financial question answering over U.S. Securities and Exchange Commission (SEC) filings requires retrieving and synthesizing heterogeneous evidence dispersed across long, standardized, and highly redundant disclosures. Existing retrieval-augmented and multi-agent systems typically derive retrieval queries directly from the user's question and rank candidates by semantic similarity. Together, these choices create prior-corpus misalignment: a mismatch between model priors and the target filings' structure, terminology, and evidence standards. As a result, query generation misses corpus-specific evidence, while semantic reranking favors topically similar but evidentially invalid false-positive chunks. We propose FinSAgent, an evidence-grounded multi-agent framework that reframes SEC filing QA as corpus-aligned retrieval planning and corrects both ends with a single principle: inject corpus-side conditioning wherever model priors would otherwise dominate. FinSAgent combines (1) role-specialized agents anchored to the mandated 10-K item structure, (2) database-aware query decomposition that conditions each agent's sub-queries on a lightweight, summary-level view of the local corpus, and (3) multi-path retrieval with a learned feature-gated reranker that separates evidential validity from semantic similarity. Across five offline financial QA benchmarks, FinSAgent improves retrieval coverage and answer correctness over strong single-agent and multi-agent baselines; in a three-arm randomized online experiment with 1,000 anonymous user ratings, it also receives higher scores than baselines. △ Less

Submitted 21 July, 2026; v1 submitted 20 July, 2026; originally announced July 2026.

Comments: 20 pages, 14 figures, 9 tables

MSC Class: H.3.3; I.2.7; I.2.11

Hidden Decoding at Scale: Latent Computation Scaling for Large Language Models

Authors: Aiwei Liu, Cheng Shi, Chuhan Wu, Ci Lei, Di Lu, Donald He, Fan Zhang, Fanhao Kong, Feifei Zhang, Guan Wang, Haicheng Wang, Haoyu Liu, Houjin Yu, Jiachen Ding, Jiayi Feng, Jie Zhou, Jijun Chi, Jindi Shi, Jing Lei, Junjie Zhang, Laiyi Li, Le Tian, Linhao Zhang, Miao Fan, Sijun Zhang , et al. (23 additional authors not shown)

Abstract: Scaling Large Language Models (LLMs) has been driven mainly by enlarging the Transformer backbone, but for an already-strong model this requires another round of costly pretraining. We study whether an existing backbone can keep improving by allocating more computation to each token while leaving the Transformer backbone fixed. Depth-recurrent (looped) Transformers pursue this goal but are hard to… ▽ More Scaling Large Language Models (LLMs) has been driven mainly by enlarging the Transformer backbone, but for an already-strong model this requires another round of costly pretraining. We study whether an existing backbone can keep improving by allocating more computation to each token while leaving the Transformer backbone fixed. Depth-recurrent (looped) Transformers pursue this goal but are hard to scale, because looped computation does not fit naturally with the pipeline parallelism used to train the largest models. We add computation along the sequence-length dimension, where the extra computation is simply a longer input and stays compatible with standard large-model training. We propose Hidden Decoding, a sequence-length scaling method applied during continued pretraining (CPT). It expands each token into n streams with independent embedding tables and keeps the intermediate streams' key-value cache as context, so each token performs more internal computation without adding or widening Transformer layers. To keep this affordable at scale, we introduce Stream-Factorized Attention, in which most layers attend only within each stream and only a few layers mix across streams, reducing the attention cost from quadratic to roughly linear in n. Experiments support two scaling results. At frontier scale, we train WeLM-HD4-80B and WeLM-HD4-617B at n=4 and improve their matched non-HD baselines, making Hidden Decoding the first demonstrated sequence-length scaling method at the 100B+ MoE scale. Across expansion factors, the gains grow as n increases, showing that sequence-length expansion is a practical fixed-backbone scaling path for frontier-scale LLMs. △ Less

Submitted 9 July, 2026; originally announced July 2026.

Comments: 30 pages, 9 figures

MSC Class: 68T50 ACM Class: I.2.7

Reductive monoids over general base

Authors: Jingren Chi, Simon Jacques

Abstract: We develop a theory of affine algebraic monoids over general base schemes whose unit groups are split reductive groups. Our main result is a classification theorem for such objects, generalizing works of Vinberg and Rittatore over a field. As applications, we obtain combinatorial descriptions and normality properties of orbit closures, prove a Steinberg-type theorem on adjoint quotients of reducti… ▽ More We develop a theory of affine algebraic monoids over general base schemes whose unit groups are split reductive groups. Our main result is a classification theorem for such objects, generalizing works of Vinberg and Rittatore over a field. As applications, we obtain combinatorial descriptions and normality properties of orbit closures, prove a Steinberg-type theorem on adjoint quotients of reductive monoids over general base schemes, and construct finite type integral models of the Vinberg monoids. A main tool in our construction is Lusztig's theory of modified quantum groups and their canonical bases. △ Less

Submitted 30 June, 2026; originally announced July 2026.

Comments: 76 pages

Equivalued affine springer fibers in mixed characteristic

Authors: Jingren Chi

Abstract: We study Witt-vector affine Springer fibers for tame equi-valued conjugacy classes in tamely ramified groups. Similar to the approach of Goresky-Kottwitz-MacPherson in the equal characteristic setting, we show that they admit pavings by perfections of iterated affine space bundles over smooth Hessenberg varieties. Along the way we prove a version of the Chevalley restriction theorem for the dual o… ▽ More We study Witt-vector affine Springer fibers for tame equi-valued conjugacy classes in tamely ramified groups. Similar to the approach of Goresky-Kottwitz-MacPherson in the equal characteristic setting, we show that they admit pavings by perfections of iterated affine space bundles over smooth Hessenberg varieties. Along the way we prove a version of the Chevalley restriction theorem for the dual of Lie algebras. △ Less

Submitted 29 June, 2026; originally announced June 2026.

Comments: 24 pages

An overview of the geometry of Kottwitz-Viehmann varieties

Authors: Jingren Chi

Abstract: This is an update of an expository article on the geometrization of orbital integrals of spherical Hecke functions on reductive groups over non-archimedean local fields, appeared in Proceedings of ICCM 2019. Compared to the published version, we add a last section on an example in SL3 case. This is an update of an expository article on the geometrization of orbital integrals of spherical Hecke functions on reductive groups over non-archimedean local fields, appeared in Proceedings of ICCM 2019. Compared to the published version, we add a last section on an example in SL3 case. △ Less

Submitted 29 June, 2026; originally announced June 2026.

Comments: 20 pages

Journal ref: Proceedings of the International Consortuim of Chinese Mathematicians 2019, vol.2, International Press of Boston, Inc, 2024, ISBN:978-157146-426-2

Muse Spark Safety & Preparedness Report

Authors: Cristina Menghini, Peter Ney, Hamza Kwisaba, Zifan, Wang, Miles Turpin, Felix Binder, Jean-Christophe Testud, Aidan Boyd, Nathaniel Li, Ivan Evtimov, Klaudia Krawiecka, Arman Zharmagambetov, Jeremy Kritz, Alexander R. Fabbri, Daniel Song, Jinpeng Miao, Joonas Hjelt, Meghna Ramani, Leona Lan, Reza Aghajani, Joanna Bitton, Mahesh Pasupuleti, Devin Norder, Khalid El-Arini , et al. (95 additional authors not shown)

Abstract: Muse Spark is the latest large language model developed by Meta. In this report, we first present evaluations for catastrophic risk domains under Meta's Advanced AI Scaling Framework, along with the evidence that informed our launch decision. We then discuss additional considerations, such as Muse Spark's broader content safety and behavioral profile, that are relevant to overall safety but fall o… ▽ More Muse Spark is the latest large language model developed by Meta. In this report, we first present evaluations for catastrophic risk domains under Meta's Advanced AI Scaling Framework, along with the evidence that informed our launch decision. We then discuss additional considerations, such as Muse Spark's broader content safety and behavioral profile, that are relevant to overall safety but fall outside the catastrophic risk domains governed by the Framework. Our preparedness results covering Chemical and Biological, Cybersecurity, and Loss of Control risks assess Muse Spark's deployment within Meta AI as presenting acceptable levels of residual risks under our Advanced AI Scaling Framework. We conducted a broad set of evaluations targeting dual-use and high-risk capabilities across these catastrophic risk domains. Those evaluations identified elevated risks prior to mitigations, with Chemical and Biological capabilities assessed as likely reaching the "high risk" category under the Advanced AI Scaling Framework before safeguards were applied. We have implemented a multi-layered set of mitigations that address the identified risks, and Muse Spark demonstrates state-of-the-art refusal across a range of benchmarks related to hazardous workflows in chemistry and biology. We therefore release Muse Spark as the underlying model of Meta AI. △ Less

Submitted 14 May, 2026; originally announced June 2026.

Comments: 159 pages, 57 figures

MMAE: A Massive Multitask Audio Editing Benchmark

Authors: Ziyang Ma, Ruiqi Yan, Ruiyang Xu, Jie Fang, Zhikang Niu, Yi-Wen Chao, Wenming Tu, Tianrui Wang, Auden, Qi Chen, Wenxi Chen, Jiaying Chi, Yanru Huo, Zixuan Jiang, Xiquan Li, Yalin Li, Junxi Liu, Minghao Liu, Binghao Qiang, Yijia Shan, Zheshu Song, Tian Tan, Zixiang Wang, Zeyu Xie, Zhifei Xie , et al. (13 additional authors not shown)

Abstract: We introduce MMAE, a Massive Multitask Audio Editing benchmark, serving as the first comprehensive evaluation testbed designed for general-purpose instruction-based audio editing. Spurred by the shift toward intelligent creation, interactive editing has rapidly expanded from visual domains, pioneered by models like Nano-banana 2 for images and Gemini-Omni for video, into audio. However, the curren… ▽ More We introduce MMAE, a Massive Multitask Audio Editing benchmark, serving as the first comprehensive evaluation testbed designed for general-purpose instruction-based audio editing. Spurred by the shift toward intelligent creation, interactive editing has rapidly expanded from visual domains, pioneered by models like Nano-banana 2 for images and Gemini-Omni for video, into audio. However, the current evaluation infrastructure lags severely, remaining highly fragmented and restricted to specific subdomains or basic operations. Unlike existing benchmarks that are limited in scope, MMAE extends to a broad spectrum of real-world scenarios, encompassing 7 distinct audio modalities, including sound, speech, music, and their mixtures. Furthermore, we establish a comprehensive taxonomy spanning 6 levels of task complexity, from basic modifications to multi-hop reasoning and multi-round editing, 2 levels of granularity, and 8 distinct operation types. Meticulously curated through human-agent collaboration, MMAE comprises 2,000 high-fidelity samples paired with a pioneering rubric-based evaluation framework. By decomposing free-form tasks into 17,741 verifiable criteria, this robust rubric-based paradigm enables a precise, multi-dimensional assessment of both instruction following and context consistency. Our extensive evaluation of leading models reveals that current systems remain far from achieving reliable edits. Strikingly, the Exact Match Rate (EMR) consistently falls below 5% and plummets to an absolute 0% in complex, mixed-modality tasks, exposing critical bottlenecks in precise execution and structural robustness. We hope MMAE will serve as a catalyst for future advances in the intelligent creation community, providing a clear diagnostic roadmap and establishing a standardized, long-lasting evaluation paradigm for next-generation audio editing systems. △ Less

Submitted 5 June, 2026; originally announced June 2026.

Comments: Open-Source at https://github.com/ddlBoJack/MMAE

Ionization energies for Rydberg $^4 \mathrm{He}$ ($1snp\,^{1,3}P$) states using the correlated B-spline basis function method

Authors: Jing Chi, Hao Fang, Yong-Hui Zhang, Li-Yan Tang, Ting-Yun Shi

Abstract: We extend the correlated B-spline basis function (C-BSBF) method to high-precision calculations of the ionization energies of helium Rydberg $n^{1,3}P$ states ($n=24$--$35$). Using a unified basis set, we evaluate nonrelativistic energies, relativistic corrections of order $mα^4$ (including finite-mass recoil), QED contributions of order $mα^5$, and partial $mα^6$ terms (singlet-triplet mixing, on… ▽ More We extend the correlated B-spline basis function (C-BSBF) method to high-precision calculations of the ionization energies of helium Rydberg $n^{1,3}P$ states ($n=24$--$35$). Using a unified basis set, we evaluate nonrelativistic energies, relativistic corrections of order $mα^4$ (including finite-mass recoil), QED contributions of order $mα^5$, and partial $mα^6$ terms (singlet-triplet mixing, one- and two-loop radiative corrections). The remaining higher-order contributions are estimated via $1/n^3$ scaling. The resulting ionization energies achieve kHz-level accuracy and are in excellent agreement with independent Hylleraas calculations, thereby providing cross-validation between two distinct theoretical approaches. From these data, the quantum-defect parameters are determined and used to extrapolate the ionization energies up to $n=102$. Combining our Rydberg ionization energies with high-precision experimental $2S \rightarrow nP$ transition frequencies yields the ionization energies for the metastable $2^{1}S$ and $2^{3}S$ states as \num{960332040.533(10)}$_\mathrm{stat}(20)_ \mathrm{sys}$ MHz and \num{1152842742.7274(53)}$_\mathrm{stat}(25)_ \mathrm{sys}$ MHz, respectively. The C-BSBF result for the $2 \, ^1 S$ state is consistent with the experimental ionization energy obtained from Rydberg-series extrapolation, while for the $2 \, ^3 S$ state the difference is 0.019(10) MHz. △ Less

Submitted 5 June, 2026; v1 submitted 3 June, 2026; originally announced June 2026.

CODEFUSE-DEBENCH: An Empirical Study on Readability, Recompilability, and Functionality

Authors: Puzhuo Liu, Yuhan Huang, Jianlei Chi, Peng Di, Yu Jiang

Abstract: Binary decompilation aims to recover binaries into high-level source code, but existing evaluations mainly rely on syntactic similarity or single-axis readability metrics, which fail to capture practical reusability. We propose a reusability-driven evaluation paradigm that measures decompiler quality along three orthogonal dimensions: readability, recompilability, and functionality. We present DEB… ▽ More Binary decompilation aims to recover binaries into high-level source code, but existing evaluations mainly rely on syntactic similarity or single-axis readability metrics, which fail to capture practical reusability. We propose a reusability-driven evaluation paradigm that measures decompiler quality along three orthogonal dimensions: readability, recompilability, and functionality. We present DEBENCH, the first automated framework for multidimensional decompilation evaluation. DEBENCH contains 240 atomic test functions, organized into 8 source files and compiled into 640 binaries. It combines LLM-as-judge readability scoring with URAF (18 sub-dimensions), iterative compile-and-repair under a fixed 50-iteration budget, and Frida-based differential dynamic tracing at the program, function, and instruction levels. We evaluate five mainstream decompilers and three repair LLMs. Our study reveals four findings. First, the reusability cliff is steep: the best decompiler-LLM pair reaches 22.3% Exact+Partial program-level behavioral overlap but only 1.2% exact stdout match, nearly 50 points below recompilability. Second, settings that maximize readability do not maximize functionality: -O3 yields the lowest readability but the highest functionality, and Clang gives lower readability than GCC but 2.6x higher functionality. Third, cross-decompiler variation at the functional level is 20x, far larger than the 1.6x cross-LLM variation, showing that progress depends more on decompiler engines than larger repair models. Fourth, failures fall into three categories: syntactic noise, type-system collapse (about 19% of repair errors), and irreversible upstream losses such as ARM64 relocation idioms and C++ ABI features. △ Less

Submitted 28 May, 2026; originally announced May 2026.

SafeRx-Agent: A Knowledge-Grounded Multi-Agent Framework for Safe and Explainable Medication Recommendation

Authors: Xinyu Wang, Hanwei Wu, Zhenghan Tai, Sicheng Lyu, Qincheng Lu, Ziyu Zhao, Jijun Chi, Jingrui Tian, Xiao-Wen Chang, Ziyang Song

Abstract: Medication recommendation predicts medications for patient visits, but existing methods still face two key challenges. At the model level, traditional drug recommendation methods only predict structured drug codes with limited evidence grounding, while LLM agents can use richer clinical context but may lack safety verification and traceability. At the task level, existing benchmarks often use broa… ▽ More Medication recommendation predicts medications for patient visits, but existing methods still face two key challenges. At the model level, traditional drug recommendation methods only predict structured drug codes with limited evidence grounding, while LLM agents can use richer clinical context but may lack safety verification and traceability. At the task level, existing benchmarks often use broad medication categories, which ignore subgroup-level safety differences and can lead to risk overestimation. We introduce the first fine-grained medication recommendation setting based on fourth-level ATC code generation. We propose Safe Prescription Agent (SafeRx-Agent), a knowledge-grounded multi-agent framework that uses patient context, external clinical knowledge, and safety verification to recommend traceable medication sets. Experimental results on MIMIC-III and MIMIC-IV datasets show that SafeRx-Agent improves fine-grained medication prediction accuracy while controlling drug interactions, contraindications, and medication set size. △ Less

Submitted 28 May, 2026; v1 submitted 27 May, 2026; originally announced May 2026.

Factorize to Generalize: Retrieval-Guided Invariant-Dynamic Decomposition for Time Series Forecasting

Authors: Jinjin Chi, Lei Feng, Lulu Zhang, Yongcheng Jing, Yiming Wang, Ximing Li, Jialie Shen, Leszek Rutkowski, Dacheng Tao

Abstract: Time series foundation models (TSFMs) have recently achieved strong zero-shot forecasting performance through large-scale pretraining and retrieval-augmented prediction. However, our empirical analysis reveals a non-trivial limitation of retrieval-based forecasting: retrieval tends to induce more oscillatory predictions, improving performance on highly fluctuating series while degrading accuracy o… ▽ More Time series foundation models (TSFMs) have recently achieved strong zero-shot forecasting performance through large-scale pretraining and retrieval-augmented prediction. However, our empirical analysis reveals a non-trivial limitation of retrieval-based forecasting: retrieval tends to induce more oscillatory predictions, improving performance on highly fluctuating series while degrading accuracy on smoother, trend-dominated ones. This suggests that retrieved information may be fused into prediction without explicitly distinguishing stable temporal structure from instance-specific variations, which can reduce robustness under distribution shifts. We propose a Retrieval-guided Invariant-Dynamic DEcomposition framework for time series forecasting. Rather than using retrieval as auxiliary predictive context, we leverage retrieved sequences as implicit samples from related environments to guide representation decomposition. Specifically, we first construct a retrieval-aware representation via attention-based aggregation, and then introduce a retrieval-guided routing mechanism to decompose it into an invariant component capturing stable shared structure and a dynamic component modeling context-dependent variations. These two components are forecast separately and fused for final prediction, enabling the model to preserve transferable patterns while remaining adaptive to evolving dynamics. We further design training objectives that encourage invariant learning and disentanglement, and provide theoretical insight showing that retrieval aggregation reduces variance and approximates invariant representation learning without explicit environment supervision. Extensive experiments demonstrate that our method consistently improves robustness under distribution shifts and outperforms existing TSFMs and retrieval-based baselines in zero-shot forecasting settings. △ Less

Submitted 24 May, 2026; originally announced May 2026.

MuMuTestUp: Mutation-based Multi-Agent Test Case Update

Authors: Dawei Tian, Jiakun Liu, Yun Peng, Yichen Zhang, Jianlei Chi, Jun Sun, Xiaohong Su

Abstract: Modern software systems evolve rapidly under CI/CD practices, where tests are critical for quality. However, substantial code changes often render existing test cases obsolete, causing pipeline disruptions, reduced productivity, and compromised quality. Recent automatic test update approaches leverage LLMs to refine test cases via execution feedback and exact-matching context retrieval, prioritizi… ▽ More Modern software systems evolve rapidly under CI/CD practices, where tests are critical for quality. However, substantial code changes often render existing test cases obsolete, causing pipeline disruptions, reduced productivity, and compromised quality. Recent automatic test update approaches leverage LLMs to refine test cases via execution feedback and exact-matching context retrieval, prioritizing executability and line coverage but suffering three limitations: (1) neglecting test assertion adequacy, weakening fault detection; (2) relying on coarse line coverage instead of specific uncovered lines/branches; (3) using exact-matching retrieval, which fails for LLM hallucinated queries. To address these, we propose MuMuTestUp, a mutation-guided multi-agent framework with three specialized agents: Mutation Analysis (strengthens assertions via surviving mutants), Coverage Analysis (generates targeted repair instructions for uncovered lines/branches), and Semantic Retrieval (handles hallucinations via semantic-similarity search). We also construct PRBENCH, a 571-sample pull-request-level dataset from 10 open-source Java projects (validated for cross-commit update scenarios). Evaluations against state-of-the-art baselines use both open-source (Deepseek-V3.2) and closed-source (GPT-4.1) LLMs. △ Less

Submitted 18 May, 2026; originally announced May 2026.

TRACER: Verifiable Generative Provenance for Multimodal Tool-Using Agents

Authors: Bihui Yu, Caijun Jia, Jing Chi, Xiaohan Liu, Yining Wang, He Bai, Yuchen Liu, Jingxuan Wei, Junnan Zhu

Abstract: Multimodal large language models increasingly solve vision-centric tasks by calling external tools for visual inspection, OCR, retrieval, calculation, and multi-step reasoning. Current tool-using agents usually expose the executed tool trajectory and the final answer, but they rarely specify which tool observation supports each generated claim. We call this missing claim-level dependency structure… ▽ More Multimodal large language models increasingly solve vision-centric tasks by calling external tools for visual inspection, OCR, retrieval, calculation, and multi-step reasoning. Current tool-using agents usually expose the executed tool trajectory and the final answer, but they rarely specify which tool observation supports each generated claim. We call this missing claim-level dependency structure the provenance gap. The gap makes tool use hard to verify and hard to optimize, because useful evidence, redundant exploration, and unsupported reasoning are mixed in the same trajectory. We introduce TRACER, a framework for verifiable generative provenance in multimodal tool-using agents. Instead of adding citations after generation, TRACER generates each answer sentence together with a structured provenance record that identifies the supporting tool turn, evidence unit, and semantic support relation. Its relation space contains Quotation, Compression, and Inference, covering direct reuse, faithful condensation, and grounded derivation. TRACER verifies each record through schema checking, tool-turn alignment, source authenticity, and relation rationality, and then converts verified provenance into traceability constraints and provenance-derived local credit for reinforcement learning. We further construct TRACE-Bench, a benchmark for sentence-level provenance reconstruction from coarse multimodal tool trajectories. On TRACE-Bench, simply adding tools often introduces noise. With Qwen3-VL-8B, TRACER reaches 78.23% answer accuracy and 95.72% summary accuracy, outperforming the strongest closed-source tool-augmented baseline by 23.80 percentage points. Compared with tool-only supervised fine-tuning, it also reduces total test-set tool calls from 4949 to 3486. These results show that reliable multimodal tool reasoning depends on provenance-aware use of observations, not on more tool calls alone. △ Less

Submitted 10 May, 2026; originally announced May 2026.

Causal Stories from Sensor Traces: Auditing Epistemic Overreach in LLM-Generated Personal Sensing Explanations

Authors: Shanshan Zhu, Han Zhang, J. Doris Chi, Subigya Nepal, Koustuv Saha

Abstract: LLMs are increasingly used to explain personal sensing data, translating traces of activity and mood into natural-language accounts of why an anomalous day may have occurred. However, such explanations can sound coherent and personally meaningful even when the underlying evidence is sparse or missing. We introduce epistemic overreach (EO) as a measure for cases where a generated explanation implie… ▽ More LLMs are increasingly used to explain personal sensing data, translating traces of activity and mood into natural-language accounts of why an anomalous day may have occurred. However, such explanations can sound coherent and personally meaningful even when the underlying evidence is sparse or missing. We introduce epistemic overreach (EO) as a measure for cases where a generated explanation implies more than the available sensing evidence can justify. To audit how often and in what forms EO occurs, we obtained anomalous-day scenarios from three longitudinal sensing datasets of college students: StudentLife, GLOBEM, and CollegeExperience. Across activity, sleep, and affect anomalies, we generated 14,922 explanations using three LLM families -- Llama, Qwen, and GPT -- under two prompting conditions: one minimally constrained prompt and another prompt explicitly instructing models to bound claims to the data. For each scenario, we varied the amount of behavioral evidence available to the model to examine whether more evidence reduces EO. We evaluated each explanation using a structured rubric, decomposing EO into the dimensions of unsupported causal attribution, unacknowledged data gaps, overconfident language, temporal inconsistency, and diagnostic inference. We find that LLMs routinely attribute anomalous days to causes without sufficient support from the data, and that this pattern replicates across datasets, anomaly types, and model families. Further, providing richer context does not reliably reduce EO; bounded prompting helps but does not eliminate it. These findings suggest that evidential grounding should be a first-order evaluation criterion for LLM-generated personal sensing explanations, alongside fluency and plausibility. We argue that personal sensing explanations require evidential discipline: systems must distinguish what is observed, what is inferred, and what remains unknown. △ Less

Submitted 8 May, 2026; originally announced May 2026.

Decomposing the Basic Abilities of Large Language Models: Mitigating Cross-Task Interference in Multi-Task Instruct-Tuning

Authors: Bing Wang, Ximing Li, Changchun Li, Jinjin Chi, Gang Niu, Masashi Sugiyama

Abstract: Recently, the prominent performance of large language models (LLMs) has been largely driven by multi-task instruct-tuning. Unfortunately, this training paradigm suffers from a key issue, named cross-task interference, due to conflicting gradients over shared parameters among different tasks. Some previous methods mitigate this issue by isolating task-specific parameters, e.g., task-specific neuron… ▽ More Recently, the prominent performance of large language models (LLMs) has been largely driven by multi-task instruct-tuning. Unfortunately, this training paradigm suffers from a key issue, named cross-task interference, due to conflicting gradients over shared parameters among different tasks. Some previous methods mitigate this issue by isolating task-specific parameters, e.g., task-specific neuron selection and mixture-of-experts. In this paper, we empirically reveal that the cross-task interference still exists for the existing solutions because of many parameters also shared by different tasks, and accordingly, we propose a novel solution, namely Basic Abilities Decomposition for multi-task Instruct-Tuning (BADIT). Specifically, we empirically find that certain parameters are consistently co-activated, and that co-activated parameters naturally organize into base groups. This motivates us to analogize that LLMs encode several orthogonal basic abilities, and that any task can be represented as a linear combination of these abilities. Accordingly, we propose BADIT that decomposes LLM parameters into orthogonal high-singular-value LoRA experts representing basic abilities, and dynamically enforces their orthogonality during training via spherical clustering of rank-1 components. We conduct extensive experiments on the SuperNI benchmark with 6 LLMs, and empirical results demonstrate that BADIT can outperform SOTA methods and mitigate the degree of cross-task interference. △ Less

Submitted 7 May, 2026; originally announced May 2026.

Comments: Accepted by ICML 2026. 25 pages, 13 figures. Code: https://github.com/wangbing1416/BADIT

AcademiClaw: When Students Set Challenges for AI Agents

Authors: Junjie Yu, Pengrui Lu, Weiye Si, Hongliang Lu, Jiabao Wu, Kaiwen Tao, Kun Wang, Lingyu Yang, Qiran Zhang, Xiuting Guo, Xuanyu Wang, Yang Wang, Yanjie Wang, Yi Yang, Zijian Hu, Ziyi Yang, Zonghan Zhou, Binghao Qiang, Borui Zhang, Chenning Li, Enchang Zhang, Feifan Chen, Feng Jian, Fengyin Sun, Hao Qiu , et al. (53 additional authors not shown)

Abstract: Benchmarks within the OpenClaw ecosystem have thus far evaluated exclusively assistant-level tasks, leaving the academic-level capabilities of OpenClaw largely unexamined. We introduce AcademiClaw, a bilingual benchmark of 80 complex, long-horizon tasks sourced directly from university students' real academic workflows -- homework, research projects, competitions, and personal projects -- that the… ▽ More Benchmarks within the OpenClaw ecosystem have thus far evaluated exclusively assistant-level tasks, leaving the academic-level capabilities of OpenClaw largely unexamined. We introduce AcademiClaw, a bilingual benchmark of 80 complex, long-horizon tasks sourced directly from university students' real academic workflows -- homework, research projects, competitions, and personal projects -- that they found current AI agents unable to solve effectively. Curated from 230 student-submitted candidates through rigorous expert review, the final task set spans 25+ professional domains, ranging from olympiad-level mathematics and linguistics problems to GPU-intensive reinforcement learning and full-stack system debugging, with 16 tasks requiring CUDA GPU execution. Each task executes in an isolated Docker sandbox and is scored on task completion by multi-dimensional rubrics combining six complementary techniques, with an independent five-category safety audit providing additional behavioral analysis. Experiments on six frontier models show that even the best achieves only a 55\% pass rate. Further analysis uncovers sharp capability boundaries across task domains, divergent behavioral strategies among models, and a disconnect between token consumption and output quality, providing fine-grained diagnostic signals beyond what aggregate metrics reveal. We hope that AcademiClaw and its open-sourced data and code can serve as a useful resource for the OpenClaw community, driving progress toward agents that are more capable and versatile across the full breadth of real-world academic demands. All data and code are available at https://github.com/GAIR-NLP/AcademiClaw. △ Less

Submitted 4 May, 2026; originally announced May 2026.

Harmful Visual Content Manipulation Matters in Misinformation Detection Under Multimedia Scenarios

Authors: Bing Wang, Ximing Li, Changchun Li, Jinjin Chi, Tianze Li, Renchu Guan, Shengsheng Wang

Abstract: Nowadays, the widespread dissemination of misinformation across numerous social media platforms has led to severe negative effects on society. To address this challenge, the automatic detection of misinformation, particularly under multimedia scenarios, has gained significant attention from both academic and industrial communities, leading to the emergence of a research task known as Multimodal Mi… ▽ More Nowadays, the widespread dissemination of misinformation across numerous social media platforms has led to severe negative effects on society. To address this challenge, the automatic detection of misinformation, particularly under multimedia scenarios, has gained significant attention from both academic and industrial communities, leading to the emergence of a research task known as Multimodal Misinformation Detection (MMD). Typically, current MMD approaches focus on capturing the semantic relationships and inconsistency between various modalities but often overlook certain critical indicators within multimodal content. Recent research has shown that manipulated features within visual content in social media articles serve as valuable clues for MMD. Meanwhile, we argue that the potential intentions behind the manipulation, e.g., harmful and harmless, also matter in MMD. Therefore, in this study, we aim to identify such multimodal misinformation by capturing two types of features: manipulation features, which represent if visual content has been manipulated, and intention features, which assess the nature of these manipulations, distinguishing between harmful and harmless intentions. Unfortunately, the manipulation and intention labels that supervise these features to be discriminative are unknown. To address this, we introduce two weakly supervised indicators as substitutes by incorporating supplementary datasets focused on image manipulation detection and framing two different classification tasks as positive and unlabeled learning issues. With this framework, we introduce an innovative MMD approach, titled Harmful Visual Content Manipulation Matters in MMD (HAVC-M4 D). Comprehensive experiments conducted on four prevalent MMD datasets indicate that HAVC-M4 D significantly and consistently enhances the performance of existing MMD methods. △ Less

Submitted 22 March, 2026; originally announced March 2026.

Spatial-TTT: Streaming Visual-based Spatial Intelligence with Test-Time Training

Authors: Fangfu Liu, Diankun Wu, Jiawei Chi, Yimo Cai, Yi-Hsin Hung, Xumin Yu, Hao Li, Han Hu, Yongming Rao, Yueqi Duan

Abstract: Humans perceive and understand real-world spaces through a stream of visual observations. Therefore, the ability to streamingly maintain and update spatial evidence from potentially unbounded video streams is essential for spatial intelligence. The core challenge is not simply longer context windows but how spatial information is selected, organized, and retained over time. In this paper, we propo… ▽ More Humans perceive and understand real-world spaces through a stream of visual observations. Therefore, the ability to streamingly maintain and update spatial evidence from potentially unbounded video streams is essential for spatial intelligence. The core challenge is not simply longer context windows but how spatial information is selected, organized, and retained over time. In this paper, we propose Spatial-TTT towards streaming visual-based spatial intelligence with test-time training (TTT), which adapts a subset of parameters (fast weights) to capture and organize spatial evidence over long-horizon scene videos. Specifically, we design a hybrid architecture and adopt large-chunk updates parallel with sliding-window attention for efficient spatial video processing. To further promote spatial awareness, we introduce a spatial-predictive mechanism applied to TTT layers with 3D spatiotemporal convolution, which encourages the model to capture geometric correspondence and temporal continuity across frames. Beyond architecture design, we construct a dataset with dense 3D spatial descriptions, which guides the model to update its fast weights to memorize and organize global 3D spatial signals in a structured manner. Extensive experiments demonstrate that Spatial-TTT improves long-horizon spatial understanding and achieves state-of-the-art performance on video spatial benchmarks. Project page: https://liuff19.github.io/Spatial-TTT. △ Less

Submitted 12 March, 2026; originally announced March 2026.

Comments: Project Page: https://liuff19.github.io/Spatial-TTT

Which Data Matter? Embedding-Based Data Selection for Speech Recognition

Authors: Zakaria Aldeneh, Skyler Seto, Maureen de Seyssel, Jie Chi, Zijin Gu, Takuya Higuchi, Jee-weon Jung, Shinji Watanabe, David Grangier, Barry-John Theobald, Tatiana Likhomanenko

Abstract: Modern ASR systems are typically trained on large-scale pseudo-labeled, in-the-wild data spanning multiple domains. While such heterogeneous data benefit generalist models designed for broad deployment, they pose challenges for specialist models targeting specific domains: specialist models lack the capacity to learn from all available data, and one must pay closer attention to addressing the mism… ▽ More Modern ASR systems are typically trained on large-scale pseudo-labeled, in-the-wild data spanning multiple domains. While such heterogeneous data benefit generalist models designed for broad deployment, they pose challenges for specialist models targeting specific domains: specialist models lack the capacity to learn from all available data, and one must pay closer attention to addressing the mismatch between training and test conditions. In this work, we study targeted data selection as a strategy to address these challenges, selecting relevant subsets from 100k hours of in-the-wild training data to optimize performance on target domains. We represent speech samples using embeddings that capture complementary characteristic--speaker attributes, phonetic content, and semantic meaning--and analyze how relevance and diversity along these axes when performing data selection affect downstream ASR performance. Our experiments with CTC-based Conformer models show that training on a strategically selected 5% subset can exceed the performance of models trained on the full dataset by up to 36.8% relative WER reduction on target domains. △ Less

Submitted 12 March, 2026; v1 submitted 5 March, 2026; originally announced March 2026.

CFG-Ctrl: Control-Based Classifier-Free Diffusion Guidance

Authors: Hanyang Wang, Yiyang Liu, Jiawei Chi, Fangfu Liu, Ran Xue, Yueqi Duan

Abstract: Classifier-Free Guidance (CFG) has emerged as a central approach for enhancing semantic alignment in flow-based diffusion models. In this paper, we explore a unified framework called CFG-Ctrl, which reinterprets CFG as a control applied to the first-order continuous-time generative flow, using the conditional-unconditional discrepancy as an error signal to adjust the velocity field. From this pers… ▽ More Classifier-Free Guidance (CFG) has emerged as a central approach for enhancing semantic alignment in flow-based diffusion models. In this paper, we explore a unified framework called CFG-Ctrl, which reinterprets CFG as a control applied to the first-order continuous-time generative flow, using the conditional-unconditional discrepancy as an error signal to adjust the velocity field. From this perspective, we summarize vanilla CFG as a proportional controller (P-control) with fixed gain, and typical follow-up variants develop extended control-law designs derived from it. However, existing methods mainly rely on linear control, inherently leading to instability, overshooting, and degraded semantic fidelity especially on large guidance scales. To address this, we introduce Sliding Mode Control CFG (SMC-CFG), which enforces the generative flow toward a rapidly convergent sliding manifold. Specifically, we define an exponential sliding mode surface over the semantic prediction error and introduce a switching control term to establish nonlinear feedback-guided correction. Moreover, we provide a Lyapunov stability analysis to theoretically support finite-time convergence. Experiments across text-to-image generation models including Stable Diffusion 3.5, Flux, and Qwen-Image demonstrate that SMC-CFG outperforms standard CFG in semantic alignment and enhances robustness across a wide range of guidance scales. Project Page: https://hanyang-21.github.io/CFG-Ctrl △ Less

Submitted 11 March, 2026; v1 submitted 3 March, 2026; originally announced March 2026.

Comments: Accepted by CVPR 2026; Project Page: https://hanyang-21.github.io/CFG-Ctrl

Large Scale High-Dimensional Reduced-Rank Linear Discriminant Analysis

Authors: Jocelyn T. Chi

Abstract: Reduced-rank linear discriminant analysis (RRLDA) is a foundational method of dimension reduction for classification that has been useful in a wide range of applications. The goal is to identify an optimal subspace to project the observations onto that simultaneously maximizes between-group variation while minimizing within-group differences. The solution is straight forward when the number of obs… ▽ More Reduced-rank linear discriminant analysis (RRLDA) is a foundational method of dimension reduction for classification that has been useful in a wide range of applications. The goal is to identify an optimal subspace to project the observations onto that simultaneously maximizes between-group variation while minimizing within-group differences. The solution is straight forward when the number of observations is greater than the number of features but computational difficulties arise in both the high-dimensional setting, where there are more features than there are observations, and when the data are very large. Many works have proposed solutions for the high-dimensional setting and frequently involve additional assumptions or tuning parameters. We propose a fast and simple iterative algorithm for both classical and high-dimensional RRLDA on large data that is free from these additional requirements and that comes with guarantees. We also explain how RRLDA-RK provides implicit regularization towards the least norm solution without explicitly incorporating penalties. We demonstrate our algorithm on real data and highlight some results. △ Less

Submitted 11 February, 2026; originally announced February 2026.

MSC Class: 15A06; 65F10; 62H30; 62H25; 62J05

ROMAN: Reward-Orchestrated Multi-Head Attention Network for Autonomous Driving System Testing

Authors: Jianlei Chi, Yuzhen Wu, Jiaxuan Hou, Xiaodong Zhang, Ming Fan, Suhui Sun, Weijun Dai, Bo Li, Jianguo Sun, Jun Sun

Abstract: Automated Driving System (ADS) acts as the brain of autonomous vehicles, responsible for their safety and efficiency. Safe deployment requires thorough testing in diverse real-world scenarios and compliance with traffic laws like speed limits, signal obedience, and right-of-way rules. Violations like running red lights or speeding pose severe safety risks. However, current testing approaches face… ▽ More Automated Driving System (ADS) acts as the brain of autonomous vehicles, responsible for their safety and efficiency. Safe deployment requires thorough testing in diverse real-world scenarios and compliance with traffic laws like speed limits, signal obedience, and right-of-way rules. Violations like running red lights or speeding pose severe safety risks. However, current testing approaches face significant challenges: limited ability to generate complex and high-risk law-breaking scenarios, and failing to account for complex interactions involving multiple vehicles and critical situations. To address these challenges, we propose ROMAN, a novel scenario generation approach for ADS testing that combines a multi-head attention network with a traffic law weighting mechanism. ROMAN is designed to generate high-risk violation scenarios to enable more thorough and targeted ADS evaluation. The multi-head attention mechanism models interactions among vehicles, traffic signals, and other factors. The traffic law weighting mechanism implements a workflow that leverages an LLM-based risk weighting module to evaluate violations based on the two dimensions of severity and occurrence. We have evaluated ROMAN by testing the Baidu Apollo ADS within the CARLA simulation platform and conducting extensive experiments to measure its performance. Experimental results demonstrate that ROMAN surpassed state-of-the-art tools ABLE and LawBreaker by achieving 7.91% higher average violation count than ABLE and 55.96% higher than LawBreaker, while also maintaining greater scenario diversity. In addition, only ROMAN successfully generated violation scenarios for every clause of the input traffic laws, enabling it to identify more high-risk violations than existing approaches. △ Less

Submitted 5 February, 2026; originally announced February 2026.

Comments: The manuscript includes 13 pages, 8 tables, and 7 figures

Disentangled Representation Learning via Flow Matching

Authors: Jinjin Chi, Taoping Liu, Mengtao Yin, Ximing Li, Yongcheng Jing, Jialie Shen, Leszek Rutkowski, Dacheng Tao

Abstract: Disentangled representation learning aims to capture the underlying explanatory factors of observed data, enabling a principled understanding of the data-generating process. Recent advances in generative modeling have introduced new paradigms for learning such representations. However, existing diffusion-based methods encourage factor independence via inductive biases, yet frequently lack strong s… ▽ More Disentangled representation learning aims to capture the underlying explanatory factors of observed data, enabling a principled understanding of the data-generating process. Recent advances in generative modeling have introduced new paradigms for learning such representations. However, existing diffusion-based methods encourage factor independence via inductive biases, yet frequently lack strong semantic alignment. In this work, we propose a flow matching-based framework for disentangled representation learning, which casts disentanglement as learning factor-conditioned flows in a compact latent space. To enforce explicit semantic alignment, we introduce a non-overlap (orthogonality) regularizer that suppresses cross-factor interference and reduces information leakage between factors. Extensive experiments across multiple datasets demonstrate consistent improvements over representative baselines, yielding higher disentanglement scores as well as improved controllability and sample fidelity. △ Less

Submitted 10 May, 2026; v1 submitted 4 February, 2026; originally announced February 2026.

Activations as Features: Probing LLMs for Generalizable Essay Scoring Representations

Authors: Jinwei Chi, Ke Wang, Yu Chen, Xuanye Lin, Qiang Xu

Abstract: Automated essay scoring (AES) is a challenging task in cross-prompt settings due to the diversity of scoring criteria. While previous studies have focused on the output of large language models (LLMs) to improve scoring accuracy, we believe activations from intermediate layers may also provide valuable information. To explore this possibility, we evaluated the discriminative power of LLMs' activat… ▽ More Automated essay scoring (AES) is a challenging task in cross-prompt settings due to the diversity of scoring criteria. While previous studies have focused on the output of large language models (LLMs) to improve scoring accuracy, we believe activations from intermediate layers may also provide valuable information. To explore this possibility, we evaluated the discriminative power of LLMs' activations in cross-prompt essay scoring task. Specifically, we used activations to fit probes and further analyzed the effects of different models and input content of LLMs on this discriminative power. By computing the directions of essays across various trait dimensions under different prompts, we analyzed the variation in evaluation perspectives of large language models concerning essay types and traits. Results show that the activations possess strong discriminative power in evaluating essay quality and that LLMs can adapt their evaluation perspectives to different traits and essay types, effectively handling the diversity of scoring criteria in cross-prompt settings. △ Less

Submitted 22 December, 2025; originally announced December 2025.

$\text{R}^2\text{R}$: A Route-to-Rerank Post-Training Framework for Multi-Domain Decoder-Only Rerankers

Authors: Xinyu Wang, Hanwei Wu, Qingchen Hu, Zhenghan Tai, Jingrui Tian, Lei Ding, Jijun Chi, Hailin He, Tung Sum Thomas Kwok, Yufei Cui, Sicheng Lyu, Muzhi Li, Mingze Li, Xinyue Yu, Ling Zhou, Peng Lu

Abstract: Decoder-only rerankers are central to Retrieval-Augmented Generation (RAG). However, generalist models miss domain-specific nuances in high-stakes fields like finance and law, and naive fine-tuning causes surface-form overfitting and catastrophic forgetting. To address this challenge, we introduce R2R, a domain-aware framework that combines dynamic expert routing with a two-stage training strategy… ▽ More Decoder-only rerankers are central to Retrieval-Augmented Generation (RAG). However, generalist models miss domain-specific nuances in high-stakes fields like finance and law, and naive fine-tuning causes surface-form overfitting and catastrophic forgetting. To address this challenge, we introduce R2R, a domain-aware framework that combines dynamic expert routing with a two-stage training strategy, Entity Abstraction for Generalization (EAG). EAG introduces a counter-shortcut mechanism by masking the most predictive surface cues, forcing the reranker to learn domain-invariant relevance patterns rather than memorizing dataset-specific entities. To efficiently activate domain experts, R2R employs a lightweight Latent Semantic Router that probes internal representations from the frozen backbone decoder to select the optimal LoRA expert per query. Extensive experiments across different reranker backbones and diverse domains (legal, medical, and financial) demonstrate that R2R consistently surpasses generalist and single-domain fine-tuned baselines. Our results confirm that R2R is a model-agnostic and modular approach to domain specialization with strong cross-domain robustness. △ Less

Submitted 25 November, 2025; originally announced November 2025.

Comments: 13 pages, including 3 figures and 3 tables

Precise ab initio calculations of $^4$He($1snp \, ^3P_J$) fine structure of high Rydberg states

Authors: Hao Fang, Jing Chi, Xiao-Qiu Qi, Yong-Hui Zhang, Li-Yan Tang, Ting-Yun Shi

Abstract: High-precision measurements of the fine-structure splittings in helium high Rydberg states have been reported, yet corresponding ab initio benchmarks for direct comparison remain unavailable. In this work, we extend the correlated B-spline basis function (C-BSBF) method to calculate the fine-structure splittings of high Rydberg states in $^4$He. The calculations include the $mα^4$- and $mα^5$-orde… ▽ More High-precision measurements of the fine-structure splittings in helium high Rydberg states have been reported, yet corresponding ab initio benchmarks for direct comparison remain unavailable. In this work, we extend the correlated B-spline basis function (C-BSBF) method to calculate the fine-structure splittings of high Rydberg states in $^4$He. The calculations include the $mα^4$- and $mα^5$-order contributions, the singlet-triplet mixing effect, and estimated spin-dependent $mα^6$-order corrections obtained using a $1/n^3$ scaling approximation. High-precision ab initio results are obtained for principal quantum numbers $n=24$-37 with kilohertz-level accuracy and further extended to $n=45$-51 by extrapolation and fitting. The theoretical results show excellent agreement with quantum-defect theory (QDT) predictions and allow direct comparison with experimental measurements. Additionally, the discrepancy observed at $n=34$ is expected to be clarified with improved experimental precision. △ Less

Submitted 31 October, 2025; originally announced October 2025.

Seed3D 1.0: From Images to High-Fidelity Simulation-Ready 3D Assets

Authors: Jiashi Feng, Xiu Li, Jing Lin, Jiahang Liu, Gaohong Liu, Weiqiang Lou, Su Ma, Guang Shi, Qinlong Wang, Jun Wang, Zhongcong Xu, Xuanyu Yi, Zihao Yu, Jianfeng Zhang, Yifan Zhu, Rui Chen, Jinxin Chi, Zixian Du, Li Han, Lixin Huang, Kaihua Jiang, Yuhan Li, Guan Luo, Shuguang Wang, Qianyi Wu , et al. (3 additional authors not shown)

Abstract: Developing embodied AI agents requires scalable training environments that balance content diversity with physics accuracy. World simulators provide such environments but face distinct limitations: video-based methods generate diverse content but lack real-time physics feedback for interactive learning, while physics-based engines provide accurate dynamics but face scalability limitations from cos… ▽ More Developing embodied AI agents requires scalable training environments that balance content diversity with physics accuracy. World simulators provide such environments but face distinct limitations: video-based methods generate diverse content but lack real-time physics feedback for interactive learning, while physics-based engines provide accurate dynamics but face scalability limitations from costly manual asset creation. We present Seed3D 1.0, a foundation model that generates simulation-ready 3D assets from single images, addressing the scalability challenge while maintaining physics rigor. Unlike existing 3D generation models, our system produces assets with accurate geometry, well-aligned textures, and realistic physically-based materials. These assets can be directly integrated into physics engines with minimal configuration, enabling deployment in robotic manipulation and simulation training. Beyond individual objects, the system scales to complete scene generation through assembling objects into coherent environments. By enabling scalable simulation-ready content creation, Seed3D 1.0 provides a foundation for advancing physics-based world simulators. Seed3D 1.0 is now available on https://console.volcengine.com/ark/region:ark+cn-beijing/experience/vision?modelId=doubao-seed3d-1-0-250928&tab=Gen3D △ Less

Submitted 22 October, 2025; originally announced October 2025.

Comments: Seed3D 1.0 Technical Report; Official Page on https://seed.bytedance.com/seed3d

VeritasFi: An Adaptable, Multi-tiered RAG Framework for Multi-modal Financial Question Answering

Authors: Zhenghan Tai, Hanwei Wu, Qingchen Hu, Jijun Chi, Hailin He, Lei Ding, Tung Sum Thomas Kwok, Bohuai Xiao, Yuchen Hua, Suyuchen Wang, Peng Lu, Muzhi Li, Yihong Wu, Liheng Ma, Jerry Huang, Jiayi Zhang, Gonghao Zhang, Chaolong Jiang, Jingrui Tian, Sicheng Lyu, Zeyu Li, Boyu Han, Fengran Mo, Xinyue Yu, Yufei Cui , et al. (2 additional authors not shown)

Abstract: Retrieval-Augmented Generation (RAG) is becoming increasingly essential for Question Answering (QA) in the financial sector, where accurate and contextually grounded insights from complex public disclosures are crucial. However, existing financial RAG systems face two significant challenges: (1) they struggle to process heterogeneous data formats, such as text, tables, and figures; and (2) they en… ▽ More Retrieval-Augmented Generation (RAG) is becoming increasingly essential for Question Answering (QA) in the financial sector, where accurate and contextually grounded insights from complex public disclosures are crucial. However, existing financial RAG systems face two significant challenges: (1) they struggle to process heterogeneous data formats, such as text, tables, and figures; and (2) they encounter difficulties in balancing general-domain applicability with company-specific adaptation. To overcome these challenges, we present VeritasFi, an innovative hybrid RAG framework that incorporates a multi-modal preprocessing pipeline alongside a cutting-edge two-stage training strategy for its re-ranking component. VeritasFi enhances financial QA through three key innovations: (1) A multi-modal preprocessing pipeline that seamlessly transforms heterogeneous data into a coherent, machine-readable format. (2) A tripartite hybrid retrieval engine that operates in parallel, combining deep multi-path retrieval over a semantically indexed document corpus, real-time data acquisition through tool utilization, and an expert-curated memory bank for high-frequency questions, ensuring comprehensive scope, accuracy, and efficiency. (3) A two-stage training strategy for the document re-ranker, which initially constructs a general, domain-specific model using anonymized data, followed by rapid fine-tuning on company-specific data for targeted applications. By integrating our proposed designs, VeritasFi presents a groundbreaking framework that greatly enhances the adaptability and robustness of financial RAG systems, providing a scalable solution for both general-domain and company-specific QA tasks. Code accompanying this work is available at https://github.com/simplew4y/VeritasFi.git. △ Less

Submitted 12 October, 2025; originally announced October 2025.

Large Reasoning Models Learn Better Alignment from Flawed Thinking

Authors: ShengYun Peng, Eric Smith, Ivan Evtimov, Song Jiang, Pin-Yu Chen, Hongyuan Zhan, Haozhu Wang, Duen Horng Chau, Mahesh Pasupuleti, Jianfeng Chi

Abstract: Large reasoning models (LRMs) "think" by generating structured chain-of-thought (CoT) before producing a final answer, yet they still lack the ability to reason critically about safety alignment and are easily biased when a flawed premise is injected into their thought process. We propose RECAP (Robust Safety Alignment via Counter-Aligned Prefilling), a principled reinforcement learning (RL) metho… ▽ More Large reasoning models (LRMs) "think" by generating structured chain-of-thought (CoT) before producing a final answer, yet they still lack the ability to reason critically about safety alignment and are easily biased when a flawed premise is injected into their thought process. We propose RECAP (Robust Safety Alignment via Counter-Aligned Prefilling), a principled reinforcement learning (RL) method for post-training that explicitly teaches models to override flawed reasoning trajectories and reroute to safe and helpful responses. RECAP trains on a mixture of synthetically generated counter-aligned CoT prefills and standard prompts, requires no additional training cost or modifications beyond vanilla reinforcement learning from human feedback (RLHF), and substantially improves safety and jailbreak robustness, reduces overrefusal, and preserves core reasoning capability -- all while maintaining inference token budget. Extensive analysis shows that RECAP-trained models engage in self-reflection more frequently and remain robust under adaptive attacks, preserving safety even after repeated attempts to override their reasoning. △ Less

Submitted 9 April, 2026; v1 submitted 1 October, 2025; originally announced October 2025.

Nonparametric Bayesian Calibration of Computer Models

Authors: Haiyi Shi, Lei Yang, Jiarui Chi, Derek Bingham, Troy Butler, Don Estep, Haonan Wang

Abstract: Combining field data and computer models is a crucial step for making inferences, predictions, and decisions for complex science and engineering systems. We formulate and analyze a nonparametric Bayesian methodology for calibrating the distribution of parameters in a computer model using field observations. Our results include establishing; a unique nonparametric Bayesian posterior corresponding t… ▽ More Combining field data and computer models is a crucial step for making inferences, predictions, and decisions for complex science and engineering systems. We formulate and analyze a nonparametric Bayesian methodology for calibrating the distribution of parameters in a computer model using field observations. Our results include establishing; a unique nonparametric Bayesian posterior corresponding to a chosen prior with an explicit formula for the posterior density; a maximum entropy property of the posterior corresponding to the uniform prior; the almost everywhere continuity of the posterior density; and a comprehensive statistical analysis of an estimator based on importance sampling. They also include establishing the well-posedness of the nonparametric Bayesian solution of the calibration problem. We illustrate the results using several examples. △ Less

Submitted 27 June, 2026; v1 submitted 26 September, 2025; originally announced September 2025.

Comments: 47 pages, 14 figures

MSC Class: Primary 62G05; 65C60 Secondary 62P30; 62P35; 60D05; 60A10

Leveraging Audio-Visual Data to Reduce the Multilingual Gap in Self-Supervised Speech Models

Authors: María Andrea Cruz Blandón, Zakaria Aldeneh, Jie Chi, Maureen de Seyssel

Abstract: Self-supervised learning (SSL) has made significant advances in speech representation learning. Models like wav2vec 2.0 and HuBERT have achieved state-of-the-art results in tasks such as speech recognition, particularly in monolingual settings. However, multilingual SSL models tend to underperform their monolingual counterparts on each individual language, especially in multilingual scenarios with… ▽ More Self-supervised learning (SSL) has made significant advances in speech representation learning. Models like wav2vec 2.0 and HuBERT have achieved state-of-the-art results in tasks such as speech recognition, particularly in monolingual settings. However, multilingual SSL models tend to underperform their monolingual counterparts on each individual language, especially in multilingual scenarios with few languages such as the bilingual setting. In this work, we investigate a novel approach to reduce this performance gap by introducing limited visual grounding into bilingual speech SSL models. Our results show that visual grounding benefits both monolingual and bilingual models, with especially pronounced gains for the latter, reducing the multilingual performance gap on zero-shot phonetic discrimination from 31.5% for audio-only models to 8.04% with grounding. △ Less

Submitted 22 September, 2025; originally announced September 2025.

Comments: 5 pages, 2 figures

Robust LLM Training Infrastructure at ByteDance

Authors: Borui Wan, Gaohong Liu, Zuquan Song, Jun Wang, Yun Zhang, Guangming Sheng, Shuguang Wang, Houmin Wei, Chenyuan Wang, Weiqiang Lou, Xi Yang, Mofan Zhang, Kaihua Jiang, Cheng Ren, Xiaoyun Zhi, Menghan Yu, Zhe Nan, Zhuolin Zheng, Baoquan Zhong, Qinlong Wang, Huan Yu, Jinxin Chi, Wang Zhang, Yuhan Li, Zixian Du , et al. (10 additional authors not shown)

Abstract: The training scale of large language models (LLMs) has reached tens of thousands of GPUs and is still continuously expanding, enabling faster learning of larger models. Accompanying the expansion of the resource scale is the prevalence of failures (CUDA error, NaN values, job hang, etc.), which poses significant challenges to training stability. Any large-scale LLM training infrastructure should s… ▽ More The training scale of large language models (LLMs) has reached tens of thousands of GPUs and is still continuously expanding, enabling faster learning of larger models. Accompanying the expansion of the resource scale is the prevalence of failures (CUDA error, NaN values, job hang, etc.), which poses significant challenges to training stability. Any large-scale LLM training infrastructure should strive for minimal training interruption, efficient fault diagnosis, and effective failure tolerance to enable highly efficient continuous training. This paper presents ByteRobust, a large-scale GPU infrastructure management system tailored for robust and stable training of LLMs. It exploits the uniqueness of LLM training process and gives top priorities to detecting and recovering failures in a routine manner. Leveraging parallelisms and characteristics of LLM training, ByteRobust enables high-capacity fault tolerance, prompt fault demarcation, and localization with an effective data-driven approach, comprehensively ensuring continuous and efficient training of LLM tasks. ByteRobust is deployed on a production GPU platform and achieves 97% ETTR for a three-month training job on 9,600 GPUs. △ Less

Submitted 20 October, 2025; v1 submitted 19 September, 2025; originally announced September 2025.

BootSeer: Analyzing and Mitigating Initialization Bottlenecks in Large-Scale LLM Training

Authors: Rui Li, Xiaoyun Zhi, Jinxin Chi, Menghan Yu, Lixin Huang, Jia Zhu, Weilun Zhang, Xing Ma, Wenjia Liu, Zhicheng Zhu, Daowen Luo, Zuquan Song, Xin Yin, Chao Xiang, Shuguang Wang, Wencong Xiao, Gene Cooperman

Abstract: Large Language Models (LLMs) have become a cornerstone of modern AI, driving breakthroughs in natural language processing and expanding into multimodal jobs involving images, audio, and video. As with most computational software, it is important to distinguish between ordinary runtime performance and startup overhead. Prior research has focused on runtime performance: improving training efficiency… ▽ More Large Language Models (LLMs) have become a cornerstone of modern AI, driving breakthroughs in natural language processing and expanding into multimodal jobs involving images, audio, and video. As with most computational software, it is important to distinguish between ordinary runtime performance and startup overhead. Prior research has focused on runtime performance: improving training efficiency and stability. This work focuses instead on the increasingly critical issue of startup overhead in training: the delay before training jobs begin execution. Startup overhead is particularly important in large, industrial-scale LLMs, where failures occur more frequently and multiple teams operate in iterative update-debug cycles. In one of our training clusters, more than 3.5% of GPU time is wasted due to startup overhead alone. In this work, we present the first in-depth characterization of LLM training startup overhead based on real production data. We analyze the components of startup cost, quantify its direct impact, and examine how it scales with job size. These insights motivate the design of Bootseer, a system-level optimization framework that addresses three primary startup bottlenecks: (a) container image loading, (b) runtime dependency installation, and (c) model checkpoint resumption. To mitigate these bottlenecks, Bootseer introduces three techniques: (a) hot block record-and-prefetch, (b) dependency snapshotting, and (c) striped HDFS-FUSE. Bootseer has been deployed in a production environment and evaluated on real LLM training workloads, demonstrating a 50% reduction in startup overhead. △ Less

Submitted 26 January, 2026; v1 submitted 16 July, 2025; originally announced July 2025.

Comments: 18 pages, 14 figures

The Why and How of Convex Clustering

Authors: Eric C. Chi, Aaron J. Molstad, Zheming Gao, Jocelyn T. Chi

Abstract: This survey reviews a clustering method based on solving a convex optimization problem. Despite the plethora of existing clustering methods, convex clustering has several uncommon features that distinguish it from prior art. The optimization problem is free of spurious local minima, and its unique global minimizer is stable with respect to all its inputs, including the data, a tuning parameter, an… ▽ More This survey reviews a clustering method based on solving a convex optimization problem. Despite the plethora of existing clustering methods, convex clustering has several uncommon features that distinguish it from prior art. The optimization problem is free of spurious local minima, and its unique global minimizer is stable with respect to all its inputs, including the data, a tuning parameter, and weight hyperparameters. Its single tuning parameter controls the number of clusters and can be chosen using standard techniques from penalized regression. We give intuition into the behavior and theory for convex clustering as well as practical guidance. We highlight important algorithms and discuss how their computational costs scale with the problem size. Finally, we highlight the breadth of its uses and flexibility to be combined and integrated with other inferential methods. △ Less

Submitted 17 September, 2025; v1 submitted 11 July, 2025; originally announced July 2025.

LangScene-X: Reconstruct Generalizable 3D Language-Embedded Scenes with TriMap Video Diffusion

Authors: Fangfu Liu, Hao Li, Jiawei Chi, Hanyang Wang, Minghui Yang, Fudong Wang, Yueqi Duan

Abstract: Recovering 3D structures with open-vocabulary scene understanding from 2D images is a fundamental but daunting task. Recent developments have achieved this by performing per-scene optimization with embedded language information. However, they heavily rely on the calibrated dense-view reconstruction paradigm, thereby suffering from severe rendering artifacts and implausible semantic synthesis when… ▽ More Recovering 3D structures with open-vocabulary scene understanding from 2D images is a fundamental but daunting task. Recent developments have achieved this by performing per-scene optimization with embedded language information. However, they heavily rely on the calibrated dense-view reconstruction paradigm, thereby suffering from severe rendering artifacts and implausible semantic synthesis when limited views are available. In this paper, we introduce a novel generative framework, coined LangScene-X, to unify and generate 3D consistent multi-modality information for reconstruction and understanding. Powered by the generative capability of creating more consistent novel observations, we can build generalizable 3D language-embedded scenes from only sparse views. Specifically, we first train a TriMap video diffusion model that can generate appearance (RGBs), geometry (normals), and semantics (segmentation maps) from sparse inputs through progressive knowledge integration. Furthermore, we propose a Language Quantized Compressor (LQC), trained on large-scale image datasets, to efficiently encode language embeddings, enabling cross-scene generalization without per-scene retraining. Finally, we reconstruct the language surface fields by aligning language information onto the surface of 3D scenes, enabling open-ended language queries. Extensive experiments on real-world data demonstrate the superiority of our LangScene-X over state-of-the-art methods in terms of quality and generalizability. Project Page: https://liuff19.github.io/LangScene-X. △ Less

Submitted 3 July, 2025; originally announced July 2025.

Comments: Project page: https://liuff19.github.io/LangScene-X

STRICT: Stress Test of Rendering Images Containing Text

Authors: Tianyu Zhang, Xinyu Wang, Lu Li, Zhenghan Tai, Jijun Chi, Jingrui Tian, Hailin He, Suyuchen Wang

Abstract: While diffusion models have revolutionized text-to-image generation with their ability to synthesize realistic and diverse scenes, they continue to struggle to generate consistent and legible text within images. This shortcoming is commonly attributed to the locality bias inherent in diffusion-based generation, which limits their ability to model long-range spatial dependencies. In this paper, we… ▽ More While diffusion models have revolutionized text-to-image generation with their ability to synthesize realistic and diverse scenes, they continue to struggle to generate consistent and legible text within images. This shortcoming is commonly attributed to the locality bias inherent in diffusion-based generation, which limits their ability to model long-range spatial dependencies. In this paper, we introduce $\textbf{STRICT}$, a benchmark designed to systematically stress-test the ability of diffusion models to render coherent and instruction-aligned text in images. Our benchmark evaluates models across multiple dimensions: (1) the maximum length of readable text that can be generated; (2) the correctness and legibility of the generated text, and (3) the ratio of not following instructions for generating text. We evaluate several state-of-the-art models, including proprietary and open-source variants, and reveal persistent limitations in long-range consistency and instruction-following capabilities. Our findings provide insights into architectural bottlenecks and motivate future research directions in multimodal generative modeling. We release our entire evaluation pipeline at https://github.com/tianyu-z/STRICT-Bench. △ Less

Submitted 14 September, 2025; v1 submitted 25 May, 2025; originally announced May 2025.

Comments: Accepted as a main conference paper at EMNLP 2025

MSC Class: 68T50 ACM Class: I.2.7; I.4.0

Discriminating Form and Meaning in Multilingual Models with Minimal-Pair ABX Tasks

Authors: Maureen de Seyssel, Jie Chi, Skyler Seto, Maartje ter Hoeve, Masha Fedzechkina, Natalie Schluter

Abstract: We introduce a set of training-free ABX-style discrimination tasks to evaluate how multilingual language models represent language identity (form) and semantic content (meaning). Inspired from speech processing, these zero-shot tasks measure whether minimal differences in representation can be reliably detected. This offers a flexible and interpretable alternative to probing. Applied to XLM-R (Con… ▽ More We introduce a set of training-free ABX-style discrimination tasks to evaluate how multilingual language models represent language identity (form) and semantic content (meaning). Inspired from speech processing, these zero-shot tasks measure whether minimal differences in representation can be reliably detected. This offers a flexible and interpretable alternative to probing. Applied to XLM-R (Conneau et al, 2020) across pretraining checkpoints and layers, we find that language discrimination declines over training and becomes concentrated in lower layers, while meaning discrimination strengthens over time and stabilizes in deeper layers. We then explore probing tasks, showing some alignment between our metrics and linguistic learning performance. Our results position ABX tasks as a lightweight framework for analyzing the structure of multilingual representations. △ Less

Submitted 19 November, 2025; v1 submitted 23 May, 2025; originally announced May 2025.

Comments: Comments: Published in EMNLP 2025. https://aclanthology.org/2025.emnlp-main.1210.pdf

Shape it Up! Restoring LLM Safety during Finetuning

Authors: ShengYun Peng, Pin-Yu Chen, Jianfeng Chi, Seongmin Lee, Duen Horng Chau

Abstract: Finetuning large language models (LLMs) enables user-specific customization but introduces critical safety risks: even a few harmful examples can compromise safety alignment. A common mitigation strategy is to update the model more strongly on examples deemed safe, while downweighting or excluding those flagged as unsafe. However, because safety context can shift within a single example, updating… ▽ More Finetuning large language models (LLMs) enables user-specific customization but introduces critical safety risks: even a few harmful examples can compromise safety alignment. A common mitigation strategy is to update the model more strongly on examples deemed safe, while downweighting or excluding those flagged as unsafe. However, because safety context can shift within a single example, updating the model equally on both harmful and harmless parts of a response is suboptimal-a coarse treatment we term static safety shaping. In contrast, we propose dynamic safety shaping (DSS), a framework that uses fine-grained safety signals to reinforce learning from safe segments of a response while suppressing unsafe content. To enable such fine-grained control during finetuning, we introduce a key insight: guardrail models, traditionally used for filtering, can be repurposed to evaluate partial responses, tracking how safety risk evolves throughout the response, segment by segment. This leads to the Safety Trajectory Assessment of Response (STAR), a token-level signal that enables shaping to operate dynamically over the training sequence. Building on this, we present STAR-DSS, guided by STAR scores, that robustly mitigates finetuning risks and delivers substantial safety improvements across diverse threats, datasets, and model families-all without compromising capability on intended tasks. We encourage future safety research to build on dynamic shaping principles for stronger mitigation against evolving finetuning risks. Our code is publicly available at https://github.com/poloclub/star-dss. △ Less

Submitted 22 December, 2025; v1 submitted 22 May, 2025; originally announced May 2025.

Comments: NeurIPS'25

Spontaneous Speech Variables for Evaluating LLMs Cognitive Plausibility

Authors: Sheng-Fu Wang, Laurent Prevot, Jou-an Chi, Ri-Sheng Huang, Shu-Kai Hsieh

Abstract: The achievements of Large Language Models in Natural Language Processing, especially for high-resource languages, call for a better understanding of their characteristics from a cognitive perspective. Researchers have attempted to evaluate artificial models by testing their ability to predict behavioral (e.g., eye-tracking fixations) and physiological (e.g., brain responses) variables during langu… ▽ More The achievements of Large Language Models in Natural Language Processing, especially for high-resource languages, call for a better understanding of their characteristics from a cognitive perspective. Researchers have attempted to evaluate artificial models by testing their ability to predict behavioral (e.g., eye-tracking fixations) and physiological (e.g., brain responses) variables during language processing (e.g., reading/listening). In this paper, we propose using spontaneous speech corpora to derive production variables (speech reductions, prosodic prominences) and applying them in a similar fashion. More precisely, we extract. We then test models trained with a standard procedure on different pretraining datasets (written, spoken, and mixed genres) for their ability to predict these two variables. Our results show that, after some fine-tuning, the models can predict these production variables well above baselines. We also observe that spoken genre training data provides more accurate predictions than written genres. These results contribute to the broader effort of using high-quality speech corpora as benchmarks for LLMs. △ Less

Submitted 22 May, 2025; originally announced May 2025.

Comments: The 14th Workshop on Cognitive Modeling and Computational Linguistics (CMCL). May 3, 2025. Collocated with NAACL 2025

Bridging the Inter-Domain Gap through Low-Level Features for Cross-Modal Medical Image Segmentation

Authors: Pengfei Lyu, Pak-Hei Yeung, Xiaosheng Yu, Jing Xia, Jianning Chi, Chengdong Wu, Jagath C. Rajapakse

Abstract: This paper addresses the task of cross-modal medical image segmentation by exploring unsupervised domain adaptation (UDA) approaches. We propose a model-agnostic UDA framework, LowBridge, which builds on a simple observation that cross-modal images share some similar low-level features (e.g., edges) as they are depicting the same structures. Specifically, we first train a generative model to recov… ▽ More This paper addresses the task of cross-modal medical image segmentation by exploring unsupervised domain adaptation (UDA) approaches. We propose a model-agnostic UDA framework, LowBridge, which builds on a simple observation that cross-modal images share some similar low-level features (e.g., edges) as they are depicting the same structures. Specifically, we first train a generative model to recover the source images from their edge features, followed by training a segmentation model on the generated source images, separately. At test time, edge features from the target images are input to the pretrained generative model to generate source-style target domain images, which are then segmented using the pretrained segmentation network. Despite its simplicity, extensive experiments on various publicly available datasets demonstrate that \proposed achieves state-of-the-art performance, outperforming eleven existing UDA approaches under different settings. Notably, further ablation studies show that \proposed is agnostic to different types of generative and segmentation models, suggesting its potential to be seamlessly plugged with the most advanced models to achieve even more outstanding results in the future. The code is available at https://github.com/JoshuaLPF/LowBridge. △ Less

Submitted 17 May, 2025; originally announced May 2025.

Comments: 11 pages, 2 figures

FinSage: A Multi-aspect RAG System for Financial Filings Question Answering

Authors: Xinyu Wang, Jijun Chi, Zhenghan Tai, Tung Sum Thomas Kwok, Muzhi Li, Zhuhong Li, Hailin He, Yuchen Hua, Peng Lu, Suyuchen Wang, Yihong Wu, Jerry Huang, Jingrui Tian, Fengran Mo, Yufei Cui, Ling Zhou

Abstract: Leveraging large language models in real-world settings often entails a need to utilize domain-specific data and tools in order to follow the complex regulations that need to be followed for acceptable use. Within financial sectors, modern enterprises increasingly rely on Retrieval-Augmented Generation (RAG) systems to address complex compliance requirements in financial document workflows. Howeve… ▽ More Leveraging large language models in real-world settings often entails a need to utilize domain-specific data and tools in order to follow the complex regulations that need to be followed for acceptable use. Within financial sectors, modern enterprises increasingly rely on Retrieval-Augmented Generation (RAG) systems to address complex compliance requirements in financial document workflows. However, existing solutions struggle to account for the inherent heterogeneity of data (e.g., text, tables, diagrams) and evolving nature of regulatory standards used in financial filings, leading to compromised accuracy in critical information extraction. We propose the FinSage framework as a solution, utilizing a multi-aspect RAG framework tailored for regulatory compliance analysis in multi-modal financial documents. FinSage introduces three innovative components: (1) a multi-modal pre-processing pipeline that unifies diverse data formats and generates chunk-level metadata summaries, (2) a multi-path sparse-dense retrieval system augmented with query expansion (HyDE) and metadata-aware semantic search, and (3) a domain-specialized re-ranking module fine-tuned via Direct Preference Optimization (DPO) to prioritize compliance-critical content. Extensive experiments demonstrate that FinSage achieves an impressive recall of 92.51% on 75 expert-curated questions derived from surpasses the best baseline method on the FinanceBench question answering datasets by 24.06% in accuracy. Moreover, FinSage has been successfully deployed as financial question-answering agent in online meetings, where it has already served more than 1,200 people. △ Less

Submitted 13 August, 2025; v1 submitted 20 April, 2025; originally announced April 2025.

Comments: Accepted at the 34th ACM International Conference on Information and Knowledge Management (CIKM2025)

Seed1.5-Thinking: Advancing Superb Reasoning Models with Reinforcement Learning

Authors: ByteDance Seed, :, Jiaze Chen, Tiantian Fan, Xin Liu, Lingjun Liu, Zhiqi Lin, Mingxuan Wang, Chengyi Wang, Xiangpeng Wei, Wenyuan Xu, Yufeng Yuan, Yu Yue, Lin Yan, Qiying Yu, Xiaochen Zuo, Chi Zhang, Ruofei Zhu, Zhecheng An, Zhihao Bai, Yu Bao, Xingyan Bin, Jiangjie Chen, Feng Chen, Hongmin Chen , et al. (249 additional authors not shown)

Abstract: We introduce Seed1.5-Thinking, capable of reasoning through thinking before responding, resulting in improved performance on a wide range of benchmarks. Seed1.5-Thinking achieves 86.7 on AIME 2024, 55.0 on Codeforces and 77.3 on GPQA, demonstrating excellent reasoning abilities in STEM and coding. Beyond reasoning tasks, the method demonstrates notable generalization across diverse domains. For in… ▽ More We introduce Seed1.5-Thinking, capable of reasoning through thinking before responding, resulting in improved performance on a wide range of benchmarks. Seed1.5-Thinking achieves 86.7 on AIME 2024, 55.0 on Codeforces and 77.3 on GPQA, demonstrating excellent reasoning abilities in STEM and coding. Beyond reasoning tasks, the method demonstrates notable generalization across diverse domains. For instance, it surpasses DeepSeek R1 by 8% in win rate on non-reasoning tasks, indicating its broader applicability. Compared to other state-of-the-art reasoning models, Seed1.5-Thinking is a Mixture-of-Experts (MoE) model with a relatively small size, featuring 20B activated and 200B total parameters. As part of our effort to assess generalized reasoning, we develop two internal benchmarks, BeyondAIME and Codeforces, both of which will be publicly released to support future research. Model trial link: https://www.volcengine.com/experience/ark. △ Less

Submitted 29 April, 2025; v1 submitted 10 April, 2025; originally announced April 2025.

ReTool: Reinforcement Learning for Strategic Tool Use in LLMs

Authors: Jiazhan Feng, Shijue Huang, Xingwei Qu, Ge Zhang, Yujia Qin, Baoquan Zhong, Chengquan Jiang, Jinxin Chi, Wanjun Zhong

Abstract: While reasoning models (e.g., DeepSeek R1) trained with reinforcement learning (RL), excel in textual reasoning, they struggle in scenarios requiring structured problem-solving, such as geometric reasoning, concise computation, or complex equation solving-areas where computational tools like code interpreters (CI) demonstrate distinct advantages. To bridge this gap, we propose ReTool, which enhanc… ▽ More While reasoning models (e.g., DeepSeek R1) trained with reinforcement learning (RL), excel in textual reasoning, they struggle in scenarios requiring structured problem-solving, such as geometric reasoning, concise computation, or complex equation solving-areas where computational tools like code interpreters (CI) demonstrate distinct advantages. To bridge this gap, we propose ReTool, which enhances long-form reasoning with tool-integrated learning, including two key features: (1) dynamic interleaving of real-time code execution within natural language reasoning processes, and (2) an automated RL paradigm that allows policy rollouts with multi-turn real-time code execution and teaches the model in learning when and how to invoke tools based on outcome feedback. ReTool employs a systematic training framework, beginning with synthetic cold-start data generation to produce code-augmented long-form reasoning traces for fine-tuning base models. Subsequent RL training leverages task outcomes as rewards to iteratively refine the model's tool use strategy, enabling autonomous discovery of optimal tool invocation patterns without human priors. Experiments on the challenging MATH Olympiad benchmark AIME demonstrate ReTool's superiority: Our 32B model achieves 67% accuracy with 400 training steps, outperforming text-based RL baseline (40% accuracy, 1080 steps) in efficiency and performance. Remarkably, ReTool-32B attains 72.5% accuracy in extended settings, surpassing OpenAI's o1-preview by 27.9%. Further analysis reveals emergent behaviors such as code self-correction, signaling an ''aha moment'' in which the model autonomously masters adaptive tool use. These findings highlight the promise of outcome-driven tool integration for advancing complex mathematical reasoning and offer new insights into hybrid neuro-symbolic systems. △ Less

Submitted 17 April, 2025; v1 submitted 15 April, 2025; originally announced April 2025.

Comments: fix typos

VideoScene: Distilling Video Diffusion Model to Generate 3D Scenes in One Step

Authors: Hanyang Wang, Fangfu Liu, Jiawei Chi, Yueqi Duan

Abstract: Recovering 3D scenes from sparse views is a challenging task due to its inherent ill-posed problem. Conventional methods have developed specialized solutions (e.g., geometry regularization or feed-forward deterministic model) to mitigate the issue. However, they still suffer from performance degradation by minimal overlap across input views with insufficient visual information. Fortunately, recent… ▽ More Recovering 3D scenes from sparse views is a challenging task due to its inherent ill-posed problem. Conventional methods have developed specialized solutions (e.g., geometry regularization or feed-forward deterministic model) to mitigate the issue. However, they still suffer from performance degradation by minimal overlap across input views with insufficient visual information. Fortunately, recent video generative models show promise in addressing this challenge as they are capable of generating video clips with plausible 3D structures. Powered by large pretrained video diffusion models, some pioneering research start to explore the potential of video generative prior and create 3D scenes from sparse views. Despite impressive improvements, they are limited by slow inference time and the lack of 3D constraint, leading to inefficiencies and reconstruction artifacts that do not align with real-world geometry structure. In this paper, we propose VideoScene to distill the video diffusion model to generate 3D scenes in one step, aiming to build an efficient and effective tool to bridge the gap from video to 3D. Specifically, we design a 3D-aware leap flow distillation strategy to leap over time-consuming redundant information and train a dynamic denoising policy network to adaptively determine the optimal leap timestep during inference. Extensive experiments demonstrate that our VideoScene achieves faster and superior 3D scene generation results than previous video diffusion models, highlighting its potential as an efficient tool for future video to 3D applications. Project Page: https://hanyang-21.github.io/VideoScene △ Less

Submitted 3 April, 2025; v1 submitted 2 April, 2025; originally announced April 2025.

Comments: Accepted by CVPR 2025; Project Page: https://hanyang-21.github.io/VideoScene