Showing 1–50 of 297 results for author: Luo, B

Benchmarking the Benchmarks: Evaluating Automated Safety Benchmarks for Small Language Models

Authors: Nyamtulla Shaik, Fengjun Li, Bo Luo

Abstract: Small Language Models (SLMs) are increasingly deployed in resource-constrained, privacy-sensitive settings, where safety and bias failures can cause security and societal risks. However, existing AI safety\slash security\slash compliance benchmarks are designed for large language models that may not transfer reliably to SLMs. We therefore ask: Can these benchmarks effectively and reliably evaluate… ▽ More Small Language Models (SLMs) are increasingly deployed in resource-constrained, privacy-sensitive settings, where safety and bias failures can cause security and societal risks. However, existing AI safety\slash security\slash compliance benchmarks are designed for large language models that may not transfer reliably to SLMs. We therefore ask: Can these benchmarks effectively and reliably evaluate SLMs? To answer this question, we conduct a large-scale assessment of the effectiveness and robustness of these automated pipelines by evaluating five widely used benchmark suites across 26 open-source SLMs under a unified judging rubric, which assigns a score of 0, 1, or 0.5 to harmful, safe, or ambiguous/irrelevant responses, respectively. Across the benchmarks, ambiguous judgments dominate and correlate with prompt complexity and model architecture, indicating that {\em LLM-centric safety benchmarks are insufficient as standalone evidence for SLM safety assessment}. In general, the ambiguity rate increases with lexical density, output perplexity, and output length and decreases with lexical sophistication, self-coherence, and reply-prompt similarity. This reveals a capability-safety confound that mixes model capability with apparent safety. Since ambiguity is prevalent, aggregate mean-score leaderboards are mathematically brittle: model rankings change significantly under reasonable ambiguity treatments, even when the underlying outputs remain unchanged. △ Less

Submitted 17 August, 2026; originally announced August 2026.

Comments: This paper is accepted for publication at ESORICS 2026

SkillTrace: Traversing a Query-Skill Graph for Composable LLM Agents

Authors: Yue Yao, Shengyuan Wang, Xin Chen, Minke Zhang, Jia He, Bingjun Luo, Tom Gedeon

Abstract: Large language model agents increasingly solve complex tasks by composing reusable skills from a library. To address this, the key challenge is not merely to retrieve individually relevant skills, but to identify a complete and executable skill composition. In this paper, we argue that this problem can be solved in a graph with three levels: compositional relations among skill queries, similarity… ▽ More Large language model agents increasingly solve complex tasks by composing reusable skills from a library. To address this, the key challenge is not merely to retrieve individually relevant skills, but to identify a complete and executable skill composition. In this paper, we argue that this problem can be solved in a graph with three levels: compositional relations among skill queries, similarity between queries and candidates in the skill library, and the dependencies among the selected candidates. We introduce SkillTrace, which organizes the user query into a semantic hierarchy, matches skill queries and candidates, and propagates over the skill dependencies. Experiments on SkillsBench and ALFWorld demonstrate that SkillTrace achieves state-of-the-art performance, reaching a success rate of 53.17% on SkillsBench and 91.43% on ALFWorld. SkillTrace also delivers consistent improvements across different backbone language models, demonstrating the generality and robustness of graph-based skill retrieval. △ Less

Submitted 4 August, 2026; v1 submitted 3 August, 2026; originally announced August 2026.

Quantum Speedups for Stochastic Optimization with Heavy-Tailed Noise

Authors: Bin Luo, Chengchang Liu, Jonathan Allcock, Shengyu Zhang, John C. S. Lui

Abstract: We study stochastic optimization with heavy-tailed gradient noise. We first propose a novel quantum mean estimator for multivariate heavy-tailed random variables that achieves lower query complexity than optimal classical estimators in the low-dimensional regime. We further develop an unbiased quantum mean estimator by applying a generalized multi-level Monte Carlo technique. We prove quantum lowe… ▽ More We study stochastic optimization with heavy-tailed gradient noise. We first propose a novel quantum mean estimator for multivariate heavy-tailed random variables that achieves lower query complexity than optimal classical estimators in the low-dimensional regime. We further develop an unbiased quantum mean estimator by applying a generalized multi-level Monte Carlo technique. We prove quantum lower bounds showing that, when the dimension $d$ of the random vector is small and can be viewed as a constant, our quantum estimators are optimal up to logarithmic factors. We further derive stronger dimension-dependent lower bounds for tail index $p>4/3$, showing that a nontrivial dependence on the dimension is unavoidable in the low-dimensional regime. Based on these estimators, we propose a quantum normalized stochastic gradient descent method ($\texttt{QNSGD}$), which finds an $ε$-stationary point using $\tilde{\mathcal{O}}\big(\sqrt d\,ε^{-\frac{5p-4}{2p-2}}\big)$ queries to the quantum stochastic gradient oracle. For a convex objective function, we propose a quantum projected stochastic gradient descent method ($\texttt{QPSGD}$), which computes a solution with $ε$-optimal solution using $\tilde{\mathcal{O}}\big(\sqrt d\,ε^{-\frac{3p-2}{2p-2}}+ε^{-2}\big)$ queries in expectation. These sharper bounds improve upon the classical lower bounds $Ω\big(ε^{-\frac{3p-2}{p-1}}\big)$ for nonconvex problems and $Ω\big(ε^{-\frac{p}{p-1}}\big)$ for convex problems in the low-dimensional regimes $d\lesssimε^{-\frac{p}{p-1}}$ and $d\lesssimε^{-\frac{2-p}{p-1}}$, respectively. △ Less

Submitted 29 July, 2026; v1 submitted 28 July, 2026; originally announced July 2026.

Comments: 56 pages

Spatio-Temporal Conditional Denoising Transformer for Modality-Missing RGBT Tracking

Authors: Andong Lu, Ziyi Zha, Jiandong Jin, Shihao Li, Chenglong Li, Jin Tang, Bin Luo

Abstract: Missing modalities in RGBT tracking often lead to incomplete and unstable multimodal feature representations that greatly degrade the performance. Existing methods typically attempt to recover missing modalities from available ones, but the quality of data generated in challenging scenarios might be unsatisfactory. In addition, current approaches exhibit limited flexibility in processing both miss… ▽ More Missing modalities in RGBT tracking often lead to incomplete and unstable multimodal feature representations that greatly degrade the performance. Existing methods typically attempt to recover missing modalities from available ones, but the quality of data generated in challenging scenarios might be unsatisfactory. In addition, current approaches exhibit limited flexibility in processing both missing and complete data. To overcome these limitations, we propose a Spatio-temporal Conditional Denoising Transformer (SCDT), which integrates the spatial cues and the temporal context to adaptively perform information reconstruction of missing modalities and feature enhancement of weak modalities in a unified framework, for robust modality-missing RGBT tracking. In particular, SCDT leverages the short-term temporal cues from recent historical frames to capture the fine-grained temporal correlations and the long-term temporal cues encoding modality evolution to capture the global context. By jointly exploiting long short-term temporal contexts as the conditions, SCDT progressively guides noisy features of available modalities to learn reliable and temporally consistent multimodal representations. Furthermore, SCDT introduces a noisemodulated adaptation mechanism that dynamically adjusts its behavior according to the modal availability, enabling a single framework to unify feature learning under both modality-missing and complete scenarios without changing the architecture or parameters. Extensive experiments on three public benchmark datasets demonstrate that our method consistently outperforms state-of-the-art methods. The code is available here. △ Less

Submitted 27 July, 2026; originally announced July 2026.

Comments: Accepted by CVPR2026

Do Agent Benchmarks Measure Capability? Protocol Validity in the Age of Agentic AI

Authors: Jiaqi Shao, Hanck Chen, Wei Zhang, Maxm Pan, Bing Luo

Abstract: Agent benchmarks increasingly evaluate repository editing, web research, terminal use, and long-horizon interaction. Their scores support capability claims only when the evaluation protocol keeps the intended capability necessary for success. Recent reward-hacking benchmarks and system reports show that agents can instead recover public solutions, read evaluation artifacts, infer generator structu… ▽ More Agent benchmarks increasingly evaluate repository editing, web research, terminal use, and long-horizon interaction. Their scores support capability claims only when the evaluation protocol keeps the intended capability necessary for success. Recent reward-hacking benchmarks and system reports show that agents can instead recover public solutions, read evaluation artifacts, infer generator structure, manipulate feedback, or benefit from invalid scoring paths; existing responses do not provide a common procedure for attributing these shortcuts and quantifying their effect across benchmarks. We formulate protocol validity and introduce HackDetect, a post-hoc audit that identifies an exposure, determines how the agent used it, and assesses whether the resulting score is misleading. We quantify score inflation with the Mislead gap, defined as the exploit score minus the intended score. We audit 2,385 traces across 15 agent benchmarks and find evidence of exposures and reward hacking in 67.0% of Frontier Science traces and 66.7% of AutoLab tasks. Across paired comparisons, we measure score inflation of 0.45-1.00, showing that benchmark reports should provide evidence that scores reflect the intended capability. △ Less

Submitted 24 July, 2026; originally announced July 2026.

PrivDNN: A Secure Multi-Party Computation Framework for Deep Learning using Partial DNN Encryption

Authors: Liangqin Ren, Zeyan Liu, Fengjun Li, Kaitai Liang, Zhu Li, Bo Luo

Abstract: In the past decade, we have witnessed an exponential growth of deep learning models, platforms, and applications. While existing DL applications and Machine Learning as a service (MLaaS) frameworks assume fully trusted models, the need for privacy-preserving DNN evaluation arises. In a secure multi-party computation scenario, both the model and the data are considered proprietary, i.e., the model… ▽ More In the past decade, we have witnessed an exponential growth of deep learning models, platforms, and applications. While existing DL applications and Machine Learning as a service (MLaaS) frameworks assume fully trusted models, the need for privacy-preserving DNN evaluation arises. In a secure multi-party computation scenario, both the model and the data are considered proprietary, i.e., the model owner does not want to reveal the highly valuable DL model to the user, while the user does not wish to disclose their private data samples either. Conventional privacy-preserving deep learning solutions ask the users to send encrypted samples to the model owners, who must handle the heavy lifting of ciphertext-domain computation with homomorphic encryption. In this paper, we present a novel solution, namely, PrivDNN, which (1) offloads the computation to the user side by sharing an encrypted deep learning model with them, (2) significantly improves the efficiency of DNN evaluation using partial DNN encryption, (3) ensures model accuracy and model privacy using a core neuron selection and encryption scheme. Experimental results show that PrivDNN reduces privacy-preserving DNN inference time and memory requirement by up to 97% while maintaining model performance and privacy. Codes can be found at https://github.com/LiangqinRen/PrivDNN △ Less

Submitted 23 July, 2026; originally announced July 2026.

Comments: Published in Proceedings on Privacy Enhancing Technologies (PoPETs 2024)

Journal ref: Proceedings on Privacy Enhancing Technologies, 2024(3), 477-494

Adversarial Style Optimization: Enhancing VLM Jailbreaks by GRPO-based Stylistic Triggers Optimization

Authors: Bingjun Luo, Jialin Guo, Yue Yao, Xinpeng Ding

Abstract: Multimodal Large Language Models (MLLMs) have achieved impressive performance, but their safety alignment remains vulnerable to jailbreak attacks. Existing content-based jailbreaks are often inconsistent and show unsatisfying performance against the rapidly evolving MLLMs, failing to exploit non-content-based vulnerabilities. Unlike previous research, we empirically find that MLLMs exhibit a Styli… ▽ More Multimodal Large Language Models (MLLMs) have achieved impressive performance, but their safety alignment remains vulnerable to jailbreak attacks. Existing content-based jailbreaks are often inconsistent and show unsatisfying performance against the rapidly evolving MLLMs, failing to exploit non-content-based vulnerabilities. Unlike previous research, we empirically find that MLLMs exhibit a Stylistic Inconsistency between their comprehension ability and safety ability: MLLMs can robustly understand content regardless of visual style, yet their defense mechanisms can be easily bypassed by specific stylistic triggers. Based on this finding, we propose Adversarial Style Optimization (ASO), a plug-and-play enhancement module to amplify existing visual jailbreaks. ASO fine-tunes an image-editing model to superimpose an optimized stylistic modification onto a given adversarial image, using a Group Relative Policy Optimization (GRPO) agent guided by a Structurally-Tiered Reward Function that combines a logit-based signal for detecting explicit refusals with a high-fidelity semantic evaluation from a powerful judge model. Extensive experiments show that ASO significantly enhances the ASR of SOTA attacks, demonstrating that stylistic biases are a scalable vector for red-teaming MLLMs. Our code is available at https://github.com/bingjunluo/ASO. △ Less

Submitted 1 June, 2026; originally announced July 2026.

Comments: Accepted by CVPR 2026 (Oral)

PhantomSeal: Proactive Deepfakes Defense with Identity/Context Protection and Forensic Tracing

Authors: Liangqin Ren, Zeyan Liu, Ye Wang, Yuxin Chen, Fengjun Li, Bo Luo

Abstract: Deepfakes, especially face-swapping attacks, pose significant challenges to authenticity, security, and ethics across science, engineering, and society. While most existing detection/tracing approaches operate post hoc, proactive defenses that aim to intervene before deepfake generation remain limited in terms of real-world effectiveness. In this paper, we present PhantomSeal, the first proactive… ▽ More Deepfakes, especially face-swapping attacks, pose significant challenges to authenticity, security, and ethics across science, engineering, and society. While most existing detection/tracing approaches operate post hoc, proactive defenses that aim to intervene before deepfake generation remain limited in terms of real-world effectiveness. In this paper, we present PhantomSeal, the first proactive defense to simultaneously protect both the identity and the context of users' images from being used in face-swapping attacks, while supporting forensic tracing. We present a novel cloaking technique that embeds a selected identity as a stealthy identifier. This mechanism steers the deepfake generation process toward producing content that resembles the chosen cloak identity, thereby preventing successful face-swapping while enabling effective feature-based forensic analysis. The effectiveness and robustness of PhantomSeal is demonstrated in extensive experiments across different face-swapping architectures and models. For example, it reduces the attack success rate of SimSwap, an advanced deepfake model, to 0.30%, and correctly identifies 97.97% of manipulated content. The source codes is available at https://github.com/LiangqinRen/PhantomSeal. △ Less

Submitted 3 August, 2026; v1 submitted 20 July, 2026; originally announced July 2026.

Comments: 22 pages, 5 figures, and 24 tables. Extended version with additional technical details and appendices. Accepted by ACM CCS 2026

Dual-Correlation Hypergraph Network for Unaligned RGBT Video Object Detection and A Large-scale Benchmark

Authors: Qishun Wang, Yapeng Li, Bin Luo, Zhengzheng Tu, Chenglong Li

Abstract: RGB-Thermal (RGBT) Video Object Detection (VOD) has gained significant traction due to its ability to overcome the limitations of conventional RGB-based VOD under challenging conditions. However, spatial misalignment commonly exists between RGBT image pairs. To address this, we propose a Dual-Correlation Hypergraph Network (DHNet) that captures high-dimensional complementary information by explici… ▽ More RGB-Thermal (RGBT) Video Object Detection (VOD) has gained significant traction due to its ability to overcome the limitations of conventional RGB-based VOD under challenging conditions. However, spatial misalignment commonly exists between RGBT image pairs. To address this, we propose a Dual-Correlation Hypergraph Network (DHNet) that captures high-dimensional complementary information by explicitly modeling two types of correlations: temporal correlation across consecutive frames and spatial correlation from cross-modal features. Specifically, we first design a Patch-based Spatial Alignment Module (PSAM) to sequentially align the multimodal features at the local region level. Subsequently, we introduce a Dual Hypergraph Fusion Module (DHFM), which constructs separate temporal and multimodal hypergraphs to enhance object discriminability through dual-correlation learning. Furthermore, the field currently lacks a large-scale, scene-diverse benchmark dataset for comprehensive evaluation. To address this gap, we construct DVT-VOD1000, a large-scale RGBT VOD dataset containing 1,000 video sequences with 103,464 RGBT image pairs. The dataset covers diverse scenarios, including campuses, parks, transportation, rural areas, night scenes, rain, and snow. Comprehensive experiments on VT-VOD50 and our DVT-VOD1000 demonstrate that DHNet achieves state-of-the-art detection accuracy. The dataset and source code will be made publicly available on https://github.com/tzz-ahu/ to support academic research. △ Less

Submitted 9 July, 2026; originally announced July 2026.

ASFR-Net: Adversarial Alignment and Spatio-Frequency Refinement Network for Heterogeneous Remote Sensing Image Change Detection

Authors: Xin-Jie Wu, Zhi-Hui You, Si-Bao Chen, Qing-Ling Shu, Xiao Wang, Jin Tang, Bin Luo

Abstract: The core challenge of heterogeneous change detection in remote sensing imagery lies in effectively decoupling genuine land-cover changes from significant modal disparities caused by distinct imaging mechanisms. These intrinsic inconsistencies are prone to introducing pseudo-changes, thereby constraining detection accuracy. To address this, we propose a novel, end-to-end adversarial spatio-frequenc… ▽ More The core challenge of heterogeneous change detection in remote sensing imagery lies in effectively decoupling genuine land-cover changes from significant modal disparities caused by distinct imaging mechanisms. These intrinsic inconsistencies are prone to introducing pseudo-changes, thereby constraining detection accuracy. To address this, we propose a novel, end-to-end adversarial spatio-frequency refinement network (ASFR-Net). Initially, a modality-invariant representation learner (MIR-Learner) guides the backbone to extract modality-invariant features, effectively bridging the primary domain gap. Subsequently, to address persistent residual modal differences, we design an innovative spatio-frequency synergistic enhancement module (SFEM), which identifies and suppresses sensor-specific noise and artifacts that are difficult to discern in the spatial domain by leveraging frequency-domain processing. Multi-level difference features are then computed from these refined representations and fed into a decoder equipped with cascaded hierarchical guided fusion module (HGFM) blocks to generate precise change maps. To alleviate the data scarcity in heterogeneous tasks, we construct and release a new high-resolution benchmark specifically focused on building changes: the visible-near-infrared heterogeneous change detection (VisNIR-HCD) dataset. It presents unique scientific challenges arising from deceptive visual similarity and non-linear spectral inversions, providing a robust platform for evaluating model generalization. Extensive experiments on VisNIR-HCD and public datasets demonstrate that ASFR-Net achieves state-of-the-art (SOTA) performance, significantly outperforming existing methods. The source code and the VisNIR-HCD dataset are publicly available at https://github.com/LuoYang2024/ASFR-Net. △ Less

Submitted 8 July, 2026; originally announced July 2026.

Attend, Transform, or Silence: Operator-Level Visual Skipping for Efficient Multimodal LLM Inference

Authors: Zhaoyang Luo, Runmin Dong, Miao Yang, Fan Wei, Yushan Lai, Bin Luo, Haohuan Fu

Abstract: Multimodal large language models (MLLMs) increasingly process long visual-token sequences, increasing the overall inference computation. Existing acceleration methods usually remove visual tokens or skip visual-token updates in entire layers, but these coarse strategies may discard fine-grained evidence or suppress useful operators together with redundant ones. In this paper, we study visual-token… ▽ More Multimodal large language models (MLLMs) increasingly process long visual-token sequences, increasing the overall inference computation. Existing acceleration methods usually remove visual tokens or skip visual-token updates in entire layers, but these coarse strategies may discard fine-grained evidence or suppress useful operators together with redundant ones. In this paper, we study visual-token computation from an answer-observable perspective and find that late visual-token updates can remain large while having little effect on answer-token representations. Motivated by this answer-silent redundancy, we decompose each Transformer layer into attention and FFN operators and show that useful visual computation is often operator-dominant and layer-dependent. We propose an operator-level visual-token skipping framework that preserves the full visual-token sequence while selectively bypassing redundant attention, FFN, or both. Experiments across three MLLM architectures and 10 VQA benchmarks show that our method achieves strong efficiency-accuracy trade-offs, reducing \textbf{33.7\%} TFLOPs on Qwen3-VL while retaining \textbf{99.5\%} of the vanilla model performance. △ Less

Submitted 30 June, 2026; originally announced June 2026.

UNICS: Multilingual Code Search via Unified Pseudocode and Contrastive Transfer Learning

Authors: Ye Fan, Jidong Ge, Chuanyi Li, Liguo Huang, Bin Luo

Abstract: While pre-trained models have achieved remarkable success in code search, their multilingual capabilities remain a major hurdle, plagued by data imbalance, cross-lingual semantic interference, and the loss of critical information from existing unified representations like Abstract Syntax Trees (ASTs) or Intermediate Representations (IRs). Furthermore, conventional contrastive learning strategies o… ▽ More While pre-trained models have achieved remarkable success in code search, their multilingual capabilities remain a major hurdle, plagued by data imbalance, cross-lingual semantic interference, and the loss of critical information from existing unified representations like Abstract Syntax Trees (ASTs) or Intermediate Representations (IRs). Furthermore, conventional contrastive learning strategies often rely on simplistic hard negative sampling while overlooking the potential of mining hard positives to learn code's intrinsic semantic invariance. To address these challenges, we introduce UNICS, a framework for multilingual code search built on a two-stage training strategy. In the first stage, UNICS is pre-trained on a novel dataset we constructed, which uses pseudo-code as a unified representation to learn a cross-lingual, algorithm-level logic that preserves full semantic fidelity. The second stage employs a multi-task transfer learning strategy that adapts this general knowledge to specific languages by decomposing code into semantic slices (e.g., API calls, function bodies) and incorporating tasks for hard positive mining and cross-lingual dynamic hard negative sampling. Experimental results demonstrate that UNICS achieves state-of-the-art performance across multiple multilingual and cross-lingual benchmarks, showcasing superior generalization and performance balance, especially in zero-shot transfer tasks to low-resource languages. △ Less

Submitted 26 June, 2026; originally announced June 2026.

Comments: Accepted to the ACM International Conference on the Foundations of Software Engineering (FSE 2026). 24.pages

Do More Agents Help? Controlled and Protocol-Aligned Evaluation of LLM Agent Workflows

Authors: Yuhang Fu, Ruishan Fang, Jiaqi Shao, Huiyu Zheng, Zhengtao Zhu, Bing Luo, Tao Lin

Abstract: Does adding more agents help an LLM workflow once compared systems share the same benchmark loader, tool access, answer contract, usage accounting, and trajectory logging? We introduce BenchAgent, an evaluation framework that places single-agent, fixed multi-agent (MAS), and evolving MAS workflows under one normalized execution and logging protocol. BenchAgent evaluates these substrate-internal wo… ▽ More Does adding more agents help an LLM workflow once compared systems share the same benchmark loader, tool access, answer contract, usage accounting, and trajectory logging? We introduce BenchAgent, an evaluation framework that places single-agent, fixed multi-agent (MAS), and evolving MAS workflows under one normalized execution and logging protocol. BenchAgent evaluates these substrate-internal workflows across ten reasoning, coding, and tool-use benchmarks with GPT-4.1, and separately reports a Protocol-Aligned External (PAE) GAIA study of a runtime-generated workflow. Under SI conditions, at most one of six tested MAS exceeds the matched single-agent anchor on benchmark-balanced average accuracy: EvoAgent lies within the Wilson one-run guidance, while the remaining five trail by 2.56-11.29 points and occupy more expensive accuracy-cost trade-offs. On the PAE GAIA snapshot, a Claude-Code-style runtime workflow reaches 66.72% overall and 69.23% on Level 3, more than 20 points above the strongest non-Claude baseline, Jarvis, a fixed MAS. △ Less

Submitted 3 June, 2026; originally announced June 2026.

Comments: https://github.com/LINs-lab/MASArena/tree/BenchAgent

ST-SimDiff: Balancing Spatiotemporal Similarity and Difference for Efficient Video Understanding with MLLMs

Authors: Bingjun Luo, Tony Wang, Chaoqi Chen, Xinpeng Ding

Abstract: Multimodal Large Language Models (MLLMs) face significant computational overhead when processing long videos due to the massive number of visual tokens required. To improve efficiency, existing methods primarily reduce redundancy by pruning or merging tokens based on importance or similarity. However, these approaches largely overlook a critical dimension of video content, i.e., changes and turnin… ▽ More Multimodal Large Language Models (MLLMs) face significant computational overhead when processing long videos due to the massive number of visual tokens required. To improve efficiency, existing methods primarily reduce redundancy by pruning or merging tokens based on importance or similarity. However, these approaches largely overlook a critical dimension of video content, i.e., changes and turning points, and they lack a collaborative model for spatio-temporal relationships. To address this, we propose a new perspective: similarity is for identifying redundancy, while difference is for capturing key events. Based on this, we designed a training-free framework named ST-SimDiff. We first construct a spatio-temporal graph from the visual tokens to uniformly model their complex associations. Subsequently, we employ a parallel dual-selection strategy: 1) similarity-based selection uses community detection to retain representative tokens, compressing static information; 2) temporal difference-based selection precisely locates content-changing points to preserve tokens that capture key dynamic shifts. This allows it to preserve both static and dynamic content with a minimal number of tokens. Extensive experiments show our method significantly outperforms state-of-the-art approaches while substantially reducing computational costs. Our code is available in https://github.com/bingjunluo/ST-SimDiff. △ Less

Submitted 21 May, 2026; originally announced May 2026.

Comments: Accepted by ICLR 2026

Enhancing Visual Token Representations for Video Large Language Models via Training-Free Spatial-Temporal Pooling and Gridding

Authors: Bingjun Luo, Tony Wang, Hanqi Chen, Xinpeng Ding

Abstract: Recent advances in Multimodal Large Language Models (MLLMs) have significantly advanced video understanding tasks, yet challenges remain in efficiently compressing visual tokens while preserving spatiotemporal interactions. Existing methods, such as LLaVA family, utilize simplistic pooling or interpolation techniques that overlook the intricate dynamics of visual tokens. To bridge this gap, we pro… ▽ More Recent advances in Multimodal Large Language Models (MLLMs) have significantly advanced video understanding tasks, yet challenges remain in efficiently compressing visual tokens while preserving spatiotemporal interactions. Existing methods, such as LLaVA family, utilize simplistic pooling or interpolation techniques that overlook the intricate dynamics of visual tokens. To bridge this gap, we propose ST-GridPool, a novel training-free visual token enhancement method designed specifically for Video LLMs. Our approach integrates Pyramid Temporal Gridding (PTG), which captures multi-grained spatiotemporal interactions through hierarchical temporal gridding, and Norm-based Spatial Pooling (NSP), which preserves high-information visual regions by leveraging the correlation between token norms and semantic richness. Extensive experiments on various benchmarks demonstrate that ST-GridPool consistently enhances performance of Video LLMs without requiring costly retraining. Our method offers an efficient and plug-and-play solution for improving visual token representations. Our code is available in https://github.com/bingjunluo/ST-GridPool. △ Less

Submitted 21 May, 2026; originally announced May 2026.

Comments: Accepted by ICLR 2026

Learning to Balance: Decoupled Siamese Diffusion Transformer for Reference-Based Remote Sensing Image Super-Resolution

Authors: Bin Luo, Runmin Dong, Zhaoyang Luo, Jinxiao Zhang, Jiyao Zhao, Fan Wei, Haohuan Fu

Abstract: Diffusion-based methods demonstrate significant potential for remote sensing image super-resolution at large scaling factors, particularly in reference-based super-resolution (RefSR), where high-resolution reference images provide critical fine-grained texture priors. However, existing methods often suffer from a trade-off between over-reliance on reference information, which leads to texture arti… ▽ More Diffusion-based methods demonstrate significant potential for remote sensing image super-resolution at large scaling factors, particularly in reference-based super-resolution (RefSR), where high-resolution reference images provide critical fine-grained texture priors. However, existing methods often suffer from a trade-off between over-reliance on reference information, which leads to texture artifacts, and under-utilization of such information, which results in insufficient detail recovery. To address these issues, we propose DS-DiT, a Decoupled Siamese Diffusion Transformer that decouples the interaction between low-resolution (LR) and reference (Ref) conditions within the attention mechanism. By allowing LR structural priors and Ref texture information to independently interact with the noisy latent, the framework effectively mitigates competition between the two conditional sources. To further compensate for the limited local modeling ability of global attention, we introduce a Patch-Level Weighting (PLW) module that adaptively modulates the fusion of conditional sources. In addition, the siamese architecture enables an inference-time autoguidance strategy that exploits the prediction discrepancy between strong and weak Ref conditions to improve generation quality without additional training. Experimental results across multiple datasets and scaling factors show that DS-DiT outperforms existing methods in both quantitative metrics and visual fidelity. △ Less

Submitted 29 June, 2026; v1 submitted 18 May, 2026; originally announced May 2026.

An Extensive Replication Study of the ABLoTS Approach for Bug Localization

Authors: Feifei Niu, Enshuo Zhang, Christoph Mayr-Dorn, Wesley Klewerton Guez Assunção, Liguo Huang, Jidong Ge, Bin Luo, Alexander Egyed

Abstract: Bug localization is the task of recommending source code locations (typically files) that contain the cause of a bug and hence need to be changed to fix the bug. Along these lines, information retrieval-based bug localization (IRBL) approaches have been adopted, which identify the most bug-prone files from the source code space. In current practice, a series of state-of-the-art IRBL techniques lev… ▽ More Bug localization is the task of recommending source code locations (typically files) that contain the cause of a bug and hence need to be changed to fix the bug. Along these lines, information retrieval-based bug localization (IRBL) approaches have been adopted, which identify the most bug-prone files from the source code space. In current practice, a series of state-of-the-art IRBL techniques leverage the combination of different components (e.g., similar reports, version history, and code structure) to achieve better performance. ABLoTS is a recently proposed approach with the core component, TraceScore, that utilizes requirements and traceability information between different issue reports (i.e., feature requests and bug reports) to identify buggy source code snippets with promising results. To evaluate the accuracy of these results and obtain additional insights into the practical applicability of ABLoTS, we conducted a replication study of this approach with the original dataset and also on two extended datasets (i.e., additional Java dataset and Python dataset). The original dataset consists of 11 open source Java projects with 8,494 bug reports. The extended Java dataset includes 16 more projects comprising 25,893 bug reports and corresponding source code commits. The extended Python dataset consists of 12 projects with 1,289 bug reports. While we find that the TraceScore component, which is the core of ABLoTS, produces comparable or even better results with the extended datasets, we also find that we cannot reproduce the ABLoTS results, as reported in its original paper, due to an overlooked side effect of incorrectly choosing a cut-off date that led to test data leaking into training data with significant effects on performance. △ Less

Submitted 12 May, 2026; originally announced May 2026.

Journal ref: Empirical Software Engineering, 2024, 29(6): 143

Equilibrium Residuals Expose Three Regimes of Matrix-Game Strategic Reasoning in Language Models

Authors: Wenhua Nie, Binhan Luo, Zijie Meng, Jyh-Shing Roger Jang, Ching-Wen Ma

Abstract: Large language models can score well on named game-theory benchmarks while failing on the same strategic computation once semantic cues are removed. We show this gap with procedurally generated zero-sum matrix games: a model that recognizes familiar games drops to 34%, 18%, and 2% success on anonymous $2{\times}2$, $3{\times}3$, and $5{\times}5$ payoff matrices. The benchmark separates semantic re… ▽ More Large language models can score well on named game-theory benchmarks while failing on the same strategic computation once semantic cues are removed. We show this gap with procedurally generated zero-sum matrix games: a model that recognizes familiar games drops to 34%, 18%, and 2% success on anonymous $2{\times}2$, $3{\times}3$, and $5{\times}5$ payoff matrices. The benchmark separates semantic recall, learned approximate Nash computation, and an output-interface bottleneck that limits scale. Training only on $2{\times}2$ and $3{\times}3$ games, supervised fine-tuning raises unseen $5{\times}5$--$7{\times}7$ success from 2% to 61%, while exploitability-reward training averages 37% with high seed variance. We prove that the exploitability residual is $2$-Lipschitz in payoff perturbations, unlike discontinuous vertex-returning LP equilibrium selectors, explaining why residual training can transfer under payoff shifts even when formatting instability limits mean performance. A dominated-action padding experiment provides causal evidence: trained models solve $3{\times}3$ games embedded in much larger matrices, while random-padded controls fail and dense $12{\times}12$ games remain near failure. Procedural evaluation is therefore necessary for measuring strategic reasoning, and residual rewards expose a real but format-limited route to approximate equilibrium computation. △ Less

Submitted 11 May, 2026; originally announced May 2026.

Identified-Set Geometry of Distributional Model Extraction under Top-$K$ Censored API Access

Authors: Wenhua Nie, ZiCheng Zhu, Jianan Wu, Binhan Luo, Haoran Zheng, Jyh-Shing Roger Jang

Abstract: Modern LLM APIs often reveal only top-$K$ logit scores and censor the remaining vocabulary. We study the per-position distribution-recovery limits of this access model. For censoring threshold $τ$, the compatible teacher distributions form an identified set whose total-variation diameter is exactly $U_K=(V-K)\exp(τ)/(Z_A+(V-K)\exp(τ))$, where $Z_A$ is the observed partition function. For KL recove… ▽ More Modern LLM APIs often reveal only top-$K$ logit scores and censor the remaining vocabulary. We study the per-position distribution-recovery limits of this access model. For censoring threshold $τ$, the compatible teacher distributions form an identified set whose total-variation diameter is exactly $U_K=(V-K)\exp(τ)/(Z_A+(V-K)\exp(τ))$, where $Z_A$ is the observed partition function. For KL recovery, we give a computable binary-endpoint lower bound and an asymptotically matching small-ambiguity upper bound, with an extension to reference-aware attackers. Experiments on a Qwen3 math-reasoning teacher reveal a layered extraction hierarchy: on-task top-$K$ distillation recovers 12% of private capability, full-logit distillation recovers 56% despite 99% KL closure, and generation-based extraction recovers 96%. Top-$K$ censoring therefore limits per-position distribution recovery but does not by itself prevent capability extraction, separating fidelity from transfer in prompt-only logit distillation. △ Less

Submitted 11 May, 2026; originally announced May 2026.

EdgeFlowerTune: Evaluating Federated LLM Fine-Tuning Under Realistic Edge System Constraints

Authors: Jiaxiang Geng, Yiyi Lu, Lunyu Zhao, Yan Gao, Nicholas D. Lane, Bing Luo

Abstract: Federated fine-tuning offers a promising paradigm for adapting large language models (LLMs) on edge devices by leveraging the rich, diverse, and continuously generated data from smartphones and IoT devices without compromising user data privacy. Such edge-side adaptation can improve model personalization, robustness, and responsiveness to local contexts. However, the practical feasibility of feder… ▽ More Federated fine-tuning offers a promising paradigm for adapting large language models (LLMs) on edge devices by leveraging the rich, diverse, and continuously generated data from smartphones and IoT devices without compromising user data privacy. Such edge-side adaptation can improve model personalization, robustness, and responsiveness to local contexts. However, the practical feasibility of federated LLM fine-tuning on real edge devices remains unclear, as most existing work focuses on cross-silo or simulation-based settings, overlooking the resource and runtime constraints that determine whether a method is deployable on real edge systems. We present EdgeFlowerTune, a deployment-oriented benchmark for federated LLM fine-tuning under realistic edge-system constraints. EdgeFlowerTune jointly evaluates model quality and system costs, including communication, wall-clock latency, memory usage, energy consumption, and robustness to dynamic edge conditions. To compare methods in terms of effectiveness, efficiency, and robustness, EdgeFlowerTune introduces three complementary protocols: Quality-under-Budget, Cost-to-Target, and Robustness. We instantiate EdgeFlowerTune as a real-device platform built on Flower and MobileFineTuner, spanning commercial Android smartphones and NVIDIA edge development boards. Our benchmark results show that accuracy-only evaluation can lead to misleading conclusions: methods with similar final quality may differ substantially in deployability once realistic system constraints are considered. EdgeFlowerTune provides a reproducible benchmark for system-aware evaluation of federated LLM fine-tuning at the edge. △ Less

Submitted 8 May, 2026; originally announced May 2026.

Comments: 30 pages, 10 figures

When Stored Evidence Stops Being Usable: Scale-Conditioned Evaluation of Agent Memory

Authors: Jiaqi Shao, Yiyi Lu, Yunzhen Zhang, Bing Luo

Abstract: Memory-agent evaluations report fixed-snapshot accuracy or retrieval quality, but these scores do not show whether evidence remains usable as irrelevant sessions (sessions not annotated as task-relevant evidence for the query) accumulate. We present a scale-conditioned evaluation protocol for agent memory under evidence-preserving growth: for each query, task evidence is held fixed while irrelevan… ▽ More Memory-agent evaluations report fixed-snapshot accuracy or retrieval quality, but these scores do not show whether evidence remains usable as irrelevant sessions (sessions not annotated as task-relevant evidence for the query) accumulate. We present a scale-conditioned evaluation protocol for agent memory under evidence-preserving growth: for each query, task evidence is held fixed while irrelevant sessions are added. The protocol logs agent--memory trajectories and reports four diagnostics: budget-compliant reliability, tail memory-call burden, failure-regime decomposition, and the usable-scale boundary where reliability falls below the target. Applied to LongMemEval and LoCoMo across flat, planar, and hierarchical memory interfaces, the protocol shows reliability loss is not a single phenomenon. On LongMemEval, HippoRAG stays within the two-call budget but loses 16--20 percentage points in budget-compliant reliability as irrelevant sessions are added; LiCoMemory's observed failures depend strongly on the agent, with Qwen3-8B exceeding the budget while Qwen3-32B and Qwen3-235B remain reliable in the tested range. The result supports a framework for making scalable-memory claims conditional on agent, interface, scale range, and interaction budget. △ Less

Submitted 8 May, 2026; originally announced May 2026.

Comments: 19 pages, 11 figures, preprint

ACM Class: I.2.7; H.3.3

Dynamic Pondering Sparsity-aware Mixture-of-Experts Transformer for Event Stream based Visual Object Tracking

Authors: Shiao Wang, Xiao Wang, Duoqing Yang, Wenhao Zhang, Bo Jiang, Lin Zhu, Yonghong Tian, Bin Luo

Abstract: Despite significant progress, RGB-based trackers remain vulnerable to challenging imaging conditions, such as low illumination and fast motion. Event cameras offer a promising alternative by asynchronously capturing pixel-wise brightness changes, providing high dynamic range and high temporal resolution. However, existing event-based trackers often neglect the intrinsic spatial sparsity and tempor… ▽ More Despite significant progress, RGB-based trackers remain vulnerable to challenging imaging conditions, such as low illumination and fast motion. Event cameras offer a promising alternative by asynchronously capturing pixel-wise brightness changes, providing high dynamic range and high temporal resolution. However, existing event-based trackers often neglect the intrinsic spatial sparsity and temporal density of event data, while relying on a single fixed temporal-window sampling strategy that is suboptimal under varying motion dynamics. In this paper, we propose an event sparsity-aware tracking framework that explicitly models event-density variations across multiple temporal scales. Specifically, the proposed framework progressively injects sparse, medium-density, and dense event search regions into a three-stage Vision Transformer backbone, enabling hierarchical multi-density feature learning. Furthermore, we introduce a sparsity-aware Mixture-of-Experts module to encourage expert specialization under different sparsity patterns, and design a dynamic pondering strategy to adaptively adjust the inference depth according to tracking difficulty. Extensive experiments on FE240hz, COESOT, and EventVOT demonstrate that the proposed approach achieves a favorable trade-off between tracking accuracy and computational efficiency. The source code will be released on https://github.com/Event-AHU/OpenEvTracking. △ Less

Submitted 7 May, 2026; originally announced May 2026.

SARU: A Shadow-Aware and Removal Unified Framework for Remote Sensing Images with New Benchmarks

Authors: Zi-Yang Bo, Wei Lu, Hongruixuan Chen, Si-Bao Chen, Bin Luo

Abstract: Shadows are a prevalent problem in remote sensing imagery (RSI), degrading visual quality and severely limiting the performance of downstream tasks like object detection and semantic segmentation. Most prior works treat shadow detection and removal as separate, cascaded tasks, which can lead to cumbersome process and error accumulation. Furthermore, many deep learning methods rely on paired shadow… ▽ More Shadows are a prevalent problem in remote sensing imagery (RSI), degrading visual quality and severely limiting the performance of downstream tasks like object detection and semantic segmentation. Most prior works treat shadow detection and removal as separate, cascaded tasks, which can lead to cumbersome process and error accumulation. Furthermore, many deep learning methods rely on paired shadow and non-shadow images for training, which are often unavailable in practice. To address these challenges, we propose Shadow-Aware and Removal Unified (SARU) Framework , a cohesive two-stage framework. First, its dual-branch detection module (DBCSF-Net) fuses multi-color space and semantic features to generate high-fidelity shadow masks, effectively distinguishing shadows from dark objects. Then, leveraging these masks, a novel, training-free physical algorithm (N$^2$SGSR) restores illumination by transferring properties from adjacent non-shadow regions within the single input image. To facilitate rigorous evaluation and foster future work, we also introduce two new benchmark datasets: the RSI Shadow Detection (RSISD) dataset and the Single-image Shadow Removal Benchmark (SiSRB). Extensive experiments on the AISD and RSISD datasets demonstrate that SARU achieves SOTA shadow detection performance. For shadow removal, our training-free N$^2$SGSR algorithm attains an average processing speed of approximately $1.3$s, which is over $10$ times faster than the SOTA MAOSD while maintains an SRI value close to 0.9 on both the AISD and SiSRB datasets, a level comparable to the advanced RS-GSSR method. By holistically integrating shadow detection and removal to mitigate error propagation and eliminating the dependency on paired training data, SARU establishes a robust, practical framework for real-world RSI analysis. The code and datasets are publicly available at: https://github.com/AeroVILab-AHU/SARU △ Less

Submitted 12 May, 2026; v1 submitted 28 April, 2026; originally announced April 2026.

Comments: Accepted by ISPRS

EviDep: Trustworthy Multimodal Depression Estimation via Disentangled Evidential Learning

Authors: Fangyuan Liu, Sirui Zhao, Zeyu Zhang, Jinyang Huang, Feng-Qi Cui, Bin Luo, Meng Li, Tong Xu, Enhong Chen

Abstract: Automated multimodal depression estimation in unconstrained environments is inherently challenged by naturalistic noise and complex behavioral variability. Prevailing deterministic methods, however, produce uncalibrated point estimates without quantifying predictive uncertainty, exposing decision-making to the risk of overconfident, untrustworthy estimates. To establish a reliable and trustworthy… ▽ More Automated multimodal depression estimation in unconstrained environments is inherently challenged by naturalistic noise and complex behavioral variability. Prevailing deterministic methods, however, produce uncalibrated point estimates without quantifying predictive uncertainty, exposing decision-making to the risk of overconfident, untrustworthy estimates. To establish a reliable and trustworthy estimation paradigm, we propose EviDep, an evidential learning framework that jointly quantifies depression severity alongside aleatoric and epistemic uncertainties via a Normal-Inverse-Gamma distribution. To ensure the integrity of the extracted behavioral evidence and prevent artificial confidence inflation during multimodal fusion, EviDep introduces two tailored mechanisms. First, addressing the temporal-frequency heterogeneity of behavioral cues, a Frequency-aware Feature Extraction module leverages a wavelet-based Mixture-of-Experts to dynamically decouple stable macro-level affective baselines from transient micro-level behavioral bursts, effectively filtering out task-irrelevant artifacts. Second, a Disentangled Evidential Learning strategy enforces explicit decorrelation of features in these purified representations. By separating the cross-modal shared consensus from modality-specific behavioral nuances before Bayesian fusion, this rigorous disentanglement strictly prevents the model from double-counting overlapping information. Extensive experiments on the AVEC 2013, AVEC 2014, DAIC-WOZ, and E-DAIC datasets confirm that EviDep achieves state-of-the-art predictive accuracy and superior uncertainty calibration, thereby delivering a trustworthy, risk-aware decision-support tool for depression estimation. △ Less

Submitted 8 May, 2026; v1 submitted 17 April, 2026; originally announced April 2026.

Building Extraction from Remote Sensing Imagery under Hazy and Low-light Conditions: Benchmark and Baseline

Authors: Feifei Sang, Wei Lu, Hongruixuan Chen, Sibao Chen, Bin Luo

Abstract: Building extraction from optical Remote Sensing (RS) imagery suffers from performance degradation under real-world hazy and low-light conditions. However, existing optical methods and benchmarks focus primarily on ideal clear-weather conditions. While SAR offers all-weather sensing, its side-looking geometry causes geometric distortions. To address these challenges, we introduce HaLoBuilding, the… ▽ More Building extraction from optical Remote Sensing (RS) imagery suffers from performance degradation under real-world hazy and low-light conditions. However, existing optical methods and benchmarks focus primarily on ideal clear-weather conditions. While SAR offers all-weather sensing, its side-looking geometry causes geometric distortions. To address these challenges, we introduce HaLoBuilding, the first optical benchmark specifically designed for building extraction under hazy and low-light conditions. By leveraging a same-scene multitemporal pairing strategy, we ensure pixel-level label alignment and high fidelity even under extreme degradation. Building upon this benchmark, we propose HaLoBuild-Net, a novel end-to-end framework for building extraction in adverse RS scenarios. At its core, we develop a Spatial-Frequency Focus Module (SFFM) to effectively mitigate meteorological interference on building features by coupling large receptive field attention with frequency-aware channel reweighting guided by stable low-frequency anchors. Additionally, a Global Multi-scale Guidance Module (GMGM) provides global semantic constraints to anchor building topologies, while a Mutual-Guided Fusion Module (MGFM) implements bidirectional semantic-spatial calibration to suppress shallow noise and sharpen weather-induced blurred boundaries. Extensive experiments demonstrate that HaLoBuild-Net significantly outperforms state-of-the-art methods and conventional cascaded restoration-segmentation paradigms on the HaLoBuilding dataset, while maintaining robust generalization on WHU, INRIA, and LoveDA datasets. The source code and datasets are publicly available at: https://github.com/AeroVILab-AHU/HaLoBuilding. △ Less

Submitted 16 April, 2026; originally announced April 2026.

Comments: 14 pages, 12 figures, 9 tables

Compositional-Degradation UAV Image Restoration: Conditional Decoupled MoE Network and A Benchmark

Authors: Jinquan Yan, Zhicheng Zhao, Zhengzheng Tu, Chenglong Li, Jin Tang, Bin Luo

Abstract: UAV images are critical for applications such as large-area mapping, infrastructure inspection, and emergency response. However, in real-world flight environments, a single image is often affected by multiple degradation factors, including rain, haze, and noise, undermining downstream task performance. Current unified restoration approaches typically rely on implicit degradation representations th… ▽ More UAV images are critical for applications such as large-area mapping, infrastructure inspection, and emergency response. However, in real-world flight environments, a single image is often affected by multiple degradation factors, including rain, haze, and noise, undermining downstream task performance. Current unified restoration approaches typically rely on implicit degradation representations that entangle multiple factors into a single condition, causing mutual interference among heterogeneous corrections. To this end, we propose DAME-Net, a Degradation-Aware Mixture-of-Experts Network that decouples explicit degradation perception from degradation-conditioned reconstruction for compositional UAV image restoration. Specifically, we design a Factor-wise Degradation Perception module(FDPM) to provide explicit per-factor degradation cues for the restoration stage through multi-label prediction with label-similarity-guided soft alignment, replacing implicit entangled conditions with interpretable and generalizable degradation descriptions. Moreover, we develop a Conditioned Decoupled MoE module(CDMM) that leverages these cues for stage-wise conditioning, spatial-frequency hybrid processing, and mask-constrained decoupled expert routing, enabling selective factor-specific correction while suppressing irrelevant interference. In addition, we construct the Multi-Degradation UAV Restoration benchmark (MDUR), the first large-scale UAV benchmark for compositional UAV image restoration, with 43 degradation configurations from single degradations to four-factor composites and standardized seen/unseen splits.Extensive experiments on MDUR demonstrate consistent improvements over representative unified restoration methods, with greater gains on unseen and higher-order composite degradations. Downstream experiments further validate benefits for UAV object detection. △ Less

Submitted 10 April, 2026; originally announced April 2026.

PA2D-MORL: Pareto Ascent Directional Decomposition based Multi-Objective Reinforcement Learning

Authors: Tianmeng Hu, Biao Luo

Abstract: Multi-objective reinforcement learning (MORL) provides an effective solution for decision-making problems involving conflicting objectives. However, achieving high-quality approximations to the Pareto policy set remains challenging, especially in complex tasks with continuous or high-dimensional state-action space. In this paper, we propose the Pareto Ascent Directional Decomposition based Multi-O… ▽ More Multi-objective reinforcement learning (MORL) provides an effective solution for decision-making problems involving conflicting objectives. However, achieving high-quality approximations to the Pareto policy set remains challenging, especially in complex tasks with continuous or high-dimensional state-action space. In this paper, we propose the Pareto Ascent Directional Decomposition based Multi-Objective Reinforcement Learning (PA2D-MORL) method, which constructs an efficient scheme for multi-objective problem decomposition and policy improvement, leading to a superior approximation of Pareto policy set. The proposed method leverages Pareto ascent direction to select the scalarization weights and computes the multi-objective policy gradient, which determines the policy optimization direction and ensures joint improvement on all objectives. Meanwhile, multiple policies are selectively optimized under an evolutionary framework to approximate the Pareto frontier from different directions. Additionally, a Pareto adaptive fine-tuning approach is applied to enhance the density and spread of the Pareto frontier approximation. Experiments on various multi-objective robot control tasks show that the proposed method clearly outperforms the current state-of-the-art algorithm in terms of both quality and stability of the outcomes. △ Less

Submitted 19 March, 2026; originally announced March 2026.

Comments: AAAI 2024

Journal ref: Proceedings of the AAAI Conference on Artificial Intelligence, 38(11), 12547-12555, 2024

RepoRepair: Leveraging Code Documentation for Repository-Level Automated Program Repair

Authors: Zhongqiang Pan, Chuanyi Li, Wenkang Zhong, Yi Feng, Bin Luo, Vincent Ng

Abstract: Automated program repair (APR) struggles to scale from isolated functions to full repositories, as it demands a global, task-aware understanding to locate necessary changes. Current methods, limited by context and reliant on shallow retrieval or costly agent iterations, falter on complex cross-file issues. To this end, we propose RepoRepair, a novel documentation-enhanced approach for repository-l… ▽ More Automated program repair (APR) struggles to scale from isolated functions to full repositories, as it demands a global, task-aware understanding to locate necessary changes. Current methods, limited by context and reliant on shallow retrieval or costly agent iterations, falter on complex cross-file issues. To this end, we propose RepoRepair, a novel documentation-enhanced approach for repository-level fault localization and program repair. Our core insight is to leverage LLMs to generate hierarchical code documentation (from functions to files) for code repositories, creating structured semantic abstractions that enable LLMs to comprehend repository-level context and dependencies. Specifically, RepoRepair first employs a text-based LLM (e.g., DeepSeek-V3) to generate file/function-level code documentation for repositories, which serves as auxiliary knowledge to guide fault localization. Subsequently, based on the fault localization results and the issue description, a powerful LLM (e.g., Claude-4) attempts to repair the identified suspicious code snippets. Evaluated on SWE-bench Lite, RepoRepair achieves a 45.7% repair rate at a low cost of $0.44 per fix. On SWE-bench Multimodal, it delivers state-of-the-art performance with a 37.1% repair rate despite a higher cost of $0.56 per fix, demonstrating robust and cost-effective performance across diverse problem domains. △ Less

Submitted 1 March, 2026; originally announced March 2026.

MO-MIX: Multi-Objective Multi-Agent Cooperative Decision-Making With Deep Reinforcement Learning

Authors: Tianmeng Hu, Biao Luo, Chunhua Yang, Tingwen Huang

Abstract: Deep reinforcement learning (RL) has been applied extensively to solve complex decision-making problems. In many real-world scenarios, tasks often have several conflicting objectives and may require multiple agents to cooperate, which are the multi-objective multi-agent decision-making problems. However, only few works have been conducted on this intersection. Existing approaches are limited to se… ▽ More Deep reinforcement learning (RL) has been applied extensively to solve complex decision-making problems. In many real-world scenarios, tasks often have several conflicting objectives and may require multiple agents to cooperate, which are the multi-objective multi-agent decision-making problems. However, only few works have been conducted on this intersection. Existing approaches are limited to separate fields and can only handle multi-agent decision-making with a single objective, or multi-objective decision-making with a single agent. In this paper, we propose MO-MIX to solve the multi-objective multi-agent reinforcement learning (MOMARL) problem. Our approach is based on the centralized training with decentralized execution (CTDE) framework. A weight vector representing preference over the objectives is fed into the decentralized agent network as a condition for local action-value function estimation, while a mixing network with parallel architecture is used to estimate the joint action-value function. In addition, an exploration guide approach is applied to improve the uniformity of the final non-dominated solutions. Experiments demonstrate that the proposed method can effectively solve the multi-objective multi-agent cooperative decision-making problem and generate an approximation of the Pareto set. Our approach not only significantly outperforms the baseline method in all four kinds of evaluation metrics, but also requires less computational cost. △ Less

Submitted 28 February, 2026; originally announced March 2026.

Comments: 15 pages, 10 figures, published in IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI)

Journal ref: IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 45, no. 10, pp. 12098-12112, Oct. 2023

DACESR: Degradation-Aware Conditional Embedding for Real-World Image Super-Resolution

Authors: Xiaoyan Lei, Wenlong Zhang, Biao Luo, Hui Liang, Weifeng Cao, Qiuting Lin

Abstract: Multimodal large models have shown excellent ability in addressing image super-resolution in real-world scenarios by leveraging language class as condition information, yet their abilities in degraded images remain limited. In this paper, we first revisit the capabilities of the Recognize Anything Model (RAM) for degraded images by calculating text similarity. We find that directly using contrasti… ▽ More Multimodal large models have shown excellent ability in addressing image super-resolution in real-world scenarios by leveraging language class as condition information, yet their abilities in degraded images remain limited. In this paper, we first revisit the capabilities of the Recognize Anything Model (RAM) for degraded images by calculating text similarity. We find that directly using contrastive learning to fine-tune RAM in the degraded space is difficult to achieve acceptable results. To address this issue, we employ a degradation selection strategy to propose a Real Embedding Extractor (REE), which achieves significant recognition performance gain on degraded image content through contrastive learning. Furthermore, we use a Conditional Feature Modulator (CFM) to incorporate the high-level information of REE for a powerful Mamba-based network, which can leverage effective pixel information to restore image textures and produce visually pleasing results. Extensive experiments demonstrate that the REE can effectively help image super-resolution networks balance fidelity and perceptual quality, highlighting the great potential of Mamba in real-world applications. The source code of this work will be made publicly available at: https://github.com/nathan66666/DACESR.git △ Less

Submitted 27 February, 2026; originally announced February 2026.

Comments: Accepted by TIP

NESTOR: A Nested MOE-based Neural Operator for Large-Scale PDE Pre-Training

Authors: Dengdi Sun, Xiaoya Zhou, Xiao Wang, Hao Si, Wanli Lyu, Jin Tang, Bin Luo

Abstract: Neural operators have emerged as an efficient paradigm for solving PDEs, overcoming the limitations of traditional numerical methods and significantly improving computational efficiency. However, due to the diversity and complexity of PDE systems, existing neural operators typically rely on a single network architecture, which limits their capacity to fully capture heterogeneous features and compl… ▽ More Neural operators have emerged as an efficient paradigm for solving PDEs, overcoming the limitations of traditional numerical methods and significantly improving computational efficiency. However, due to the diversity and complexity of PDE systems, existing neural operators typically rely on a single network architecture, which limits their capacity to fully capture heterogeneous features and complex system dependencies. This constraint poses a bottleneck for large-scale PDE pre-training based on neural operators. To address these challenges, we propose a large-scale PDE pre-trained neural operator based on a nested Mixture-of-Experts (MoE) framework. In particular, the image-level MoE is designed to capture global dependencies, while the token-level Sub-MoE focuses on local dependencies. Our model can selectively activate the most suitable expert networks for a given input, thereby enhancing generalization and transferability. We conduct large-scale pre-training on twelve PDE datasets from diverse sources and successfully transfer the model to downstream tasks. Extensive experiments demonstrate the effectiveness of our approach. △ Less

Submitted 25 February, 2026; originally announced February 2026.

Comments: Accepted by CVPR 2026

Robust and Efficient Tool Orchestration via Layered Execution Structures with Reflective Correction

Authors: Tao Zhe, Haoyu Wang, Bo Luo, Min Wu, Wei Fan, Xiao Luo, Zijun Yao, Haifeng Chen, Dongjie Wang

Abstract: Tool invocation is a core capability of agentic systems, yet failures often arise not from individual tool calls but from how multiple tools are organized and executed together. Existing approaches tightly couple tool execution with stepwise language reasoning or explicit planning, leading to brittle behavior and high execution overhead. To overcome these limitations, we revisit tool invocation fr… ▽ More Tool invocation is a core capability of agentic systems, yet failures often arise not from individual tool calls but from how multiple tools are organized and executed together. Existing approaches tightly couple tool execution with stepwise language reasoning or explicit planning, leading to brittle behavior and high execution overhead. To overcome these limitations, we revisit tool invocation from the perspective of tool orchestration. Our key insight is that effective orchestration does not require precise dependency graphs or fine-grained planning. Instead, a coarse-grained layer structure suffices to provide global guidance, while execution-time errors can be corrected locally. Specifically, we model tool orchestration as learning a layered execution structure that captures high-level tool dependencies, inducing layer-wise execution through context constraints. To handle execution-time failures, we introduce a schema-aware reflective correction mechanism that detects and repairs errors locally. This design confines errors to individual tool calls and avoids re-planning entire execution trajectories. This structured execution paradigm enables a lightweight and reusable orchestration component for agentic systems. Experimental results show that our approach achieves robust tool execution while reducing execution complexity and overhead. Code will be made publicly available. △ Less

Submitted 27 February, 2026; v1 submitted 21 February, 2026; originally announced February 2026.

In-Context Planning with Latent Temporal Abstractions

Authors: Baiting Luo, Yunuo Zhang, Nathaniel S. Keplinger, Samir Gupta, Abhishek Dubey, Ayan Mukhopadhyay

Abstract: Planning-based reinforcement learning for continuous control is bottlenecked by two practical issues: planning at primitive time scales leads to prohibitive branching and long horizons, while real environments are frequently partially observable and exhibit regime shifts that invalidate stationary, fully observed dynamics assumptions. We introduce I-TAP (In-Context Latent Temporal-Abstraction Plan… ▽ More Planning-based reinforcement learning for continuous control is bottlenecked by two practical issues: planning at primitive time scales leads to prohibitive branching and long horizons, while real environments are frequently partially observable and exhibit regime shifts that invalidate stationary, fully observed dynamics assumptions. We introduce I-TAP (In-Context Latent Temporal-Abstraction Planner), an offline RL framework that unifies in-context adaptation with online planning in a learned discrete temporal-abstraction space. From offline trajectories, I-TAP learns an observation-conditioned residual-quantization VAE that compresses each observation-macro-action segment into a coarse-to-fine stack of discrete residual tokens, and a temporal Transformer that autoregressively predicts these token stacks from a short recent history. The resulting sequence model acts simultaneously as a context-conditioned prior over abstract actions and a latent dynamics model. At test time, I-TAP performs Monte Carlo Tree Search directly in token space, using short histories for implicit adaptation without gradient update, and decodes selected token stacks into executable actions. Across deterministic MuJoCo, stochastic MuJoCo with per-episode latent dynamics regimes, and high-dimensional Adroit manipulation, including partially observable variants, I-TAP consistently matches or outperforms strong model-free and model-based offline baselines, demonstrating efficient and robust in-context planning under stochastic dynamics and partial observability. △ Less

Submitted 20 February, 2026; originally announced February 2026.

RIDER: 3D RNA Inverse Design with Reinforcement Learning-Guided Diffusion

Authors: Tianmeng Hu, Yongzheng Cui, Biao Luo, Ke Li

Abstract: The inverse design of RNA three-dimensional (3D) structures is crucial for engineering functional RNAs in synthetic biology and therapeutics. While recent deep learning approaches have advanced this field, they are typically optimized and evaluated using native sequence recovery, which is a limited surrogate for structural fidelity, since different sequences can fold into similar 3D structures and… ▽ More The inverse design of RNA three-dimensional (3D) structures is crucial for engineering functional RNAs in synthetic biology and therapeutics. While recent deep learning approaches have advanced this field, they are typically optimized and evaluated using native sequence recovery, which is a limited surrogate for structural fidelity, since different sequences can fold into similar 3D structures and high recovery does not necessarily indicate correct folding. To address this limitation, we propose RIDER, an RNA Inverse DEsign framework with Reinforcement learning that directly optimizes for 3D structural similarity. First, we develop and pre-train a GNN-based generative diffusion model conditioned on the target 3D structure, achieving a 9% improvement in native sequence recovery over state-of-the-art methods. Then, we fine-tune the model with an improved policy gradient algorithm using four task-specific reward functions based on 3D self-consistency metrics. Experimental results show that RIDER improves structural similarity by over 100% across all metrics and discovers designs that are distinct from native sequences. △ Less

Submitted 7 May, 2026; v1 submitted 18 February, 2026; originally announced February 2026.

Comments: Accepted as a conference paper at ICLR 2026

Generalizable IoT Traffic Representations for Cross-Network Device Identification

Authors: Arunan Sivanathan, David Warren, Deepak Mishra, Sushmita Ruj, Natasha Fernandes, Quan Z. Sheng, Minh Tran, Ben Luo, Daniel Coscia, Gustavo Batista, Hassan Habibi Gharakaheili

Abstract: Machine learning models have demonstrated strong performance in classifying network traffic and identifying Internet-of-Things (IoT) devices, enabling operators to discover and manage IoT assets at scale. However, many existing approaches rely on end-to-end supervised pipelines or task-specific fine-tuning, resulting in traffic representations that are tightly coupled to labeled datasets and deplo… ▽ More Machine learning models have demonstrated strong performance in classifying network traffic and identifying Internet-of-Things (IoT) devices, enabling operators to discover and manage IoT assets at scale. However, many existing approaches rely on end-to-end supervised pipelines or task-specific fine-tuning, resulting in traffic representations that are tightly coupled to labeled datasets and deployment environments, which can limit generalizability. In this paper, we study the problem of learning generalizable traffic representations for IoT device identification. We design compact encoder architectures that learn per-flow embeddings from unlabeled IoT traffic and evaluate them using a frozen-encoder protocol with a simple supervised classifier. Our specific contributions are threefold. (1) We develop unsupervised encoder--decoder models that learn compact traffic representations from unlabeled IoT network flows and assess their quality through reconstruction-based analysis. (2) We show that these learned representations can be used effectively for IoT device-type classification using simple, lightweight classifiers trained on frozen embeddings. (3) We provide a systematic benchmarking study against the state-of-the-art pretrained traffic encoders, showing that larger models do not necessarily yield more robust representations for IoT traffic. Using more than 18 million real IoT traffic flows collected across multiple years and deployment environments, we learn traffic representations from unlabeled data and evaluate device-type classification on disjoint labeled subsets, achieving macro F1-scores exceeding 0.9 for device-type classification and demonstrating robustness under cross-environment deployment. △ Less

Submitted 27 January, 2026; originally announced January 2026.

Comments: 15 pages, 15 figures

MSC Class: 68M12; 68T05 ACM Class: C.2.3; I.2.6; I.5.1

FTDMamba: Frequency-Assisted Temporal Dilation Mamba for Unmanned Aerial Vehicle Video Anomaly Detection

Authors: Cheng-Zhuang Liu, Si-Bao Chen, Qing-Ling Shu, Chris Ding, Jin Tang, Bin Luo

Abstract: Recent advances in video anomaly detection (VAD) mainly focus on ground-based surveillance or unmanned aerial vehicle (UAV) videos with static backgrounds, whereas research on UAV videos with dynamic backgrounds remains limited. Unlike static scenarios, dynamically captured UAV videos exhibit multi-source motion coupling, where the motion of objects and UAV-induced global motion are intricately in… ▽ More Recent advances in video anomaly detection (VAD) mainly focus on ground-based surveillance or unmanned aerial vehicle (UAV) videos with static backgrounds, whereas research on UAV videos with dynamic backgrounds remains limited. Unlike static scenarios, dynamically captured UAV videos exhibit multi-source motion coupling, where the motion of objects and UAV-induced global motion are intricately intertwined. Consequently, existing methods may misclassify normal UAV movements as anomalies or fail to capture true anomalies concealed within dynamic backgrounds. Moreover, many approaches do not adequately address the joint modeling of inter-frame continuity and local spatial correlations across diverse temporal scales. To overcome these limitations, we propose the Frequency-Assisted Temporal Dilation Mamba (FTDMamba) network for UAV VAD, including two core components: (1) a Frequency Decoupled Spatiotemporal Correlation Module, which disentangles coupled motion patterns and models global spatiotemporal dependencies through frequency analysis; and (2) a Temporal Dilation Mamba Module, which leverages Mamba's sequence modeling capability to jointly learn fine-grained temporal dynamics and local spatial structures across multiple temporal receptive fields. Additionally, unlike existing UAV VAD datasets which focus on static backgrounds, we construct a large-scale Moving UAV VAD dataset (MUVAD), comprising 222,736 frames with 240 anomaly events across 12 anomaly types. Extensive experiments demonstrate that FTDMamba achieves state-of-the-art (SOTA) performance on two public static benchmarks and the new MUVAD dataset. The code and MUVAD dataset will be available at: https://github.com/uavano/FTDMamba. △ Less

Submitted 16 January, 2026; originally announced January 2026.

CoCoPlan: Adaptive Coordination and Communication for Multi-robot Systems in Dynamic and Unknown Environments

Authors: Xintong Zhang, Junfeng Chen, Yuxiao Zhu, Bing Luo, Meng Guo

Abstract: Multi-robot systems can greatly enhance efficiency through coordination and collaboration, yet in practice, full-time communication is rarely available and interactions are constrained to close-range exchanges. Existing methods either maintain all-time connectivity, rely on fixed schedules, or adopt pairwise protocols, but none adapt effectively to dynamic spatio-temporal task distributions under… ▽ More Multi-robot systems can greatly enhance efficiency through coordination and collaboration, yet in practice, full-time communication is rarely available and interactions are constrained to close-range exchanges. Existing methods either maintain all-time connectivity, rely on fixed schedules, or adopt pairwise protocols, but none adapt effectively to dynamic spatio-temporal task distributions under limited communication, resulting in suboptimal coordination. To address this gap, we propose CoCoPlan, a unified framework that co-optimizes collaborative task planning and team-wise intermittent communication. Our approach integrates a branch-and-bound architecture that jointly encodes task assignments and communication events, an adaptive objective function that balances task efficiency against communication latency, and a communication event optimization module that strategically determines when, where and how the global connectivity should be re-established. Extensive experiments demonstrate that it outperforms state-of-the-art methods by achieving a 22.4% higher task completion rate, reducing communication overhead by 58.6%, and improving the scalability by supporting up to 100 robots in dynamic environments. Hardware experiments include the complex 2D office environment and large-scale 3D disaster-response scenario. △ Less

Submitted 15 January, 2026; originally announced January 2026.

Comments: 8 pages, 8 figures, published to RA-L

Journal ref: IEEE Robotics and Automation Letters, vol. 11, issue 3, pp. 3270-3277, March 2026

Enhancing Portfolio Optimization with Deep Learning Insights

Authors: Brandon Luo, Jim Skufca

Abstract: Our work focuses on deep learning (DL) portfolio optimization, tackling challenges in long-only, multi-asset strategies across market cycles. We propose training models with limited regime data using pre-training techniques and leveraging transformer architectures for state variable inclusion. Evaluating our approach against traditional methods shows promising results, demonstrating our models' re… ▽ More Our work focuses on deep learning (DL) portfolio optimization, tackling challenges in long-only, multi-asset strategies across market cycles. We propose training models with limited regime data using pre-training techniques and leveraging transformer architectures for state variable inclusion. Evaluating our approach against traditional methods shows promising results, demonstrating our models' resilience in volatile markets. These findings emphasize the evolving landscape of DL-driven portfolio optimization, stressing the need for adaptive strategies to navigate dynamic market conditions and improve predictive accuracy. △ Less

Submitted 12 January, 2026; originally announced January 2026.

StriderSPD: Structure-Guided Joint Representation Learning for Binary Security Patch Detection

Authors: Qingyuan Li, Chenchen Yu, Chuanyi Li, Xin-Cheng Wen, Cheryl Lee, Cuiyun Gao, Bin Luo

Abstract: Vulnerabilities severely threaten software systems, making the timely application of security patches crucial for mitigating attacks. However, software vendors often silently patch vulnerabilities with limited disclosure, where Security Patch Detection (SPD) comes to protect software assets. Recently, most SPD studies have targeted Open-Source Software (OSS), yet a large portion of real-world soft… ▽ More Vulnerabilities severely threaten software systems, making the timely application of security patches crucial for mitigating attacks. However, software vendors often silently patch vulnerabilities with limited disclosure, where Security Patch Detection (SPD) comes to protect software assets. Recently, most SPD studies have targeted Open-Source Software (OSS), yet a large portion of real-world software is closed-source, where patches are distributed as binaries without accessible source code. The limited binary SPD approaches often lift binaries to abstraction levels, i.e., assembly code or pseudo-code. However, assembly code is register-based instructions conveying limited semantics, while pseudo-code lacks parser-compatible grammar to extract structure, both hindering accurate vulnerability-fix representation learning. In addition, previous studies often obtain training and testing data from the same project for evaluation, which fails to reflect closed-source conditions. To alleviate the above challenges, we propose \textbf{\textit{StriderSPD}}, a \underline{Str}ucture-gu\underline{ide}d joint \underline{r}epresentation \underline{SPD} framework of binary code that integrates a graph branch into a large language model (LLM), leveraging structural information to guide the LLM in identifying security patches. Our novel design of the adapters in the graph branch effectively aligns the representations between assembly code and pseudo-code at the LLM's token level. We further present a two-stage training strategy to address the optimization imbalance caused by the large parameter disparity between StriderSPD's two branches, which enables proper branch fitting. To enable more realistic evaluation, we construct a binary SPD benchmark that is disjoint from prior datasets in both projects and domains and extensively evaluate StriderSPD on this benchmark. △ Less

Submitted 9 January, 2026; originally announced January 2026.

Mechanism Design for Federated Learning with Non-Monotonic Network Effects

Authors: Xiang Li, Bing Luo, Jianwei Huang, Yuan Luo

Abstract: Mechanism design is pivotal to federated learning (FL) for maximizing social welfare by coordinating self-interested clients. Existing mechanisms, however, often overlook the network effects of client participation and the diverse model performance requirements (i.e., generalization error) across applications, leading to suboptimal incentives and social welfare, or even inapplicability in real dep… ▽ More Mechanism design is pivotal to federated learning (FL) for maximizing social welfare by coordinating self-interested clients. Existing mechanisms, however, often overlook the network effects of client participation and the diverse model performance requirements (i.e., generalization error) across applications, leading to suboptimal incentives and social welfare, or even inapplicability in real deployments. To address this gap, we explore incentive mechanism design for FL with network effects and application-specific requirements of model performance. We develop a theoretical model to quantify the impact of network effects on heterogeneous client participation, revealing the non-monotonic nature of such effects. Based on these insights, we propose a Model Trading and Sharing (MoTS) framework, which enables clients to obtain FL models through either participation or purchase. To further address clients' strategic behaviors, we design a Social Welfare maximization with Application-aware and Network effects (SWAN) mechanism, exploiting model customer payments for incentivization. Experimental results on a hardware prototype demonstrate that our SWAN mechanism outperforms existing FL mechanisms, improving social welfare by up to $352.42\%$ and reducing extra incentive costs by $93.07\%$. △ Less

Submitted 8 January, 2026; originally announced January 2026.

Comments: Journal extension of Mobihoc conference version, under review of IEEE TMC

Unlocking the Pre-Trained Model as a Dual-Alignment Calibrator for Post-Trained LLMs

Authors: Beier Luo, Cheng Wang, Hongxin Wei, Sharon Li, Xuefeng Du

Abstract: Post-training improves large language models (LLMs) but often worsens confidence calibration, leading to systematic overconfidence. Recent unsupervised post-hoc methods for post-trained LMs (PoLMs) mitigate this by aligning PoLM confidence to that of well-calibrated pre-trained counterparts. However, framing calibration as static output-distribution matching overlooks the inference-time dynamics i… ▽ More Post-training improves large language models (LLMs) but often worsens confidence calibration, leading to systematic overconfidence. Recent unsupervised post-hoc methods for post-trained LMs (PoLMs) mitigate this by aligning PoLM confidence to that of well-calibrated pre-trained counterparts. However, framing calibration as static output-distribution matching overlooks the inference-time dynamics introduced by post-training. In particular, we show that calibration errors arise from two regimes: (i) confidence drift, where final confidence inflates despite largely consistent intermediate decision processes, and (ii) process drift, where intermediate inference pathways diverge. Guided by this diagnosis, we propose Dual-Align, an unsupervised post-hoc framework for dual alignment in confidence calibration. Dual-Align performs confidence alignment to correct confidence drift via final-distribution matching, and introduces process alignment to address process drift by locating the layer where trajectories diverge and realigning the stability of subsequent inference. This dual strategy learns a single temperature parameter that corrects both drift types without sacrificing post-training performance gains. Experiments show consistent improvements over baselines, reducing calibration errors and approaching a supervised oracle. △ Less

Submitted 7 January, 2026; originally announced January 2026.

CaveAgent: Transforming LLMs into Stateful Runtime Operators

Authors: Maohao Ran, Zhenglin Wan, Cooper Lin, Yanting Zhang, Hongyu Xin, Hongwei Fan, Yibo Xu, Beier Luo, Yaxin Zhou, Wangbo Zhao, Lijie Yang, Lang Feng, Fuchao Yang, Jingxuan Wu, Yiqiao Huang, Chendong Ma, Yusen Huang, Dailing Jiang, Jianbo Deng, Sirui Han, Yang You, Bo An, Yike Guo, Jun Song

Abstract: LLM-based agents are increasingly capable of complex task execution, yet current agentic systems remain constrained by text-centric paradigms that struggle with long-horizon tasks due to fragile multi-turn dependencies and context drift. We present CaveAgent, a framework that shifts tool use from ``LLM-as-Text-Generator'' to ``LLM-as-Runtime-Operator.'' CaveAgent introduces a dual-stream architect… ▽ More LLM-based agents are increasingly capable of complex task execution, yet current agentic systems remain constrained by text-centric paradigms that struggle with long-horizon tasks due to fragile multi-turn dependencies and context drift. We present CaveAgent, a framework that shifts tool use from ``LLM-as-Text-Generator'' to ``LLM-as-Runtime-Operator.'' CaveAgent introduces a dual-stream architecture that inverts the conventional paradigm: rather than treating the LLM's text context as the primary workspace with tools as auxiliary, CaveAgent elevates the persistent Python runtime as the central locus of state, with a lightweight semantic stream serving as its orchestrator. Beyond leveraging code generation to resolve interdependent sub-tasks (e.g., loops, conditionals) in a single step, CaveAgent introduces \textit{Stateful Runtime Management}: it injects, manipulates, and retrieves complex Python objects (e.g., DataFrames, database connections) that persist across turns, unlike existing code-based approaches that remain text-bound. CaveAgent further provides a runtime-integrated skill management system that extends the Agent Skills open standard, enabling ecosystem interoperability through executable skill injections. This persistence mechanism serves as a high-fidelity external memory that reduces context drift in multi-turn interactions and preserves processed data for downstream applications without information loss. Evaluations show consistent improvement across challenging benchmarks, enabling CaveAgent to handle data scales that cause context overflow in both JSON-based and code-based agents. The accessible runtime state further provides programmatically verifiable feedback, enabling automated evaluation and reward signal generation without human annotation and establishing a structural foundation for future research in Reinforcement Learning with Verifiable Rewards (RLVR). △ Less

Submitted 28 June, 2026; v1 submitted 4 January, 2026; originally announced January 2026.

Comments: ver.2

SLEI3D: Simultaneous Exploration and Inspection via Heterogeneous Fleets under Limited Communication

Authors: Junfeng Chen, Yuxiao Zhu, Xintong Zhang, Bing Luo, Meng Guo

Abstract: Robotic fleets such as unmanned aerial and ground vehicles have been widely used for routine inspections of static environments, where the areas of interest are known and planned in advance. However, in many applications, such areas of interest are unknown and should be identified online during exploration. Thus, this paper considers the problem of simultaneous exploration, inspection of unknown e… ▽ More Robotic fleets such as unmanned aerial and ground vehicles have been widely used for routine inspections of static environments, where the areas of interest are known and planned in advance. However, in many applications, such areas of interest are unknown and should be identified online during exploration. Thus, this paper considers the problem of simultaneous exploration, inspection of unknown environments and then real-time communication to a mobile ground control station to report the findings. The heterogeneous robots are equipped with different sensors, e.g., long-range lidars for fast exploration and close-range cameras for detailed inspection. Furthermore, global communication is often unavailable in such environments, where the robots can only communicate with each other via ad-hoc wireless networks when they are in close proximity and free of obstruction. This work proposes a novel planning and coordination framework (SLEI3D) that integrates the online strategies for collaborative 3D exploration, adaptive inspection and timely communication (via the intermit-tent or proactive protocols). To account for uncertainties w.r.t. the number and location of features, a multi-layer and multi-rate planning mechanism is developed for inter-and-intra robot subgroups, to actively meet and coordinate their local plans. The proposed framework is validated extensively via high-fidelity simulations of numerous large-scale missions with up to 48 robots and 384 thousand cubic meters. Hardware experiments of 7 robots are also conducted. Project website is available at https://junfengchen-robotics.github.io/SLEI3D/. △ Less

Submitted 31 December, 2025; originally announced January 2026.

RS-Prune: Training-Free Data Pruning at High Ratios for Efficient Remote Sensing Diffusion Foundation Models

Authors: Fan Wei, Runmin Dong, Yushan Lai, Yixiang Yang, Zhaoyang Luo, Jinxiao Zhang, Miao Yang, Shuai Yuan, Jiyao Zhao, Bin Luo, Haohuan Fu

Abstract: Diffusion-based remote sensing (RS) generative foundation models are cruial for downstream tasks. However, these models rely on large amounts of globally representative data, which often contain redundancy, noise, and class imbalance, reducing training efficiency and preventing convergence. Existing RS diffusion foundation models typically aggregate multiple classification datasets or apply simpli… ▽ More Diffusion-based remote sensing (RS) generative foundation models are cruial for downstream tasks. However, these models rely on large amounts of globally representative data, which often contain redundancy, noise, and class imbalance, reducing training efficiency and preventing convergence. Existing RS diffusion foundation models typically aggregate multiple classification datasets or apply simplistic deduplication, overlooking the distributional requirements of generation modeling and the heterogeneity of RS imagery. To address these limitations, we propose a training-free, two-stage data pruning approach that quickly select a high-quality subset under high pruning ratios, enabling a preliminary foundation model to converge rapidly and serve as a versatile backbone for generation, downstream fine-tuning, and other applications. Our method jointly considers local information content with global scene-level diversity and representativeness. First, an entropy-based criterion efficiently removes low-information samples. Next, leveraging RS scene classification datasets as reference benchmarks, we perform scene-aware clustering with stratified sampling to improve clustering effectiveness while reducing computational costs on large-scale unlabeled data. Finally, by balancing cluster-level uniformity and sample representativeness, the method enables fine-grained selection under high pruning ratios while preserving overall diversity and representativeness. Experiments show that, even after pruning 85\% of the training data, our method significantly improves convergence and generation quality. Furthermore, diffusion foundation models trained with our method consistently achieve state-of-the-art performance across downstream tasks, including super-resolution and semantic image synthesis. This data pruning paradigm offers practical guidance for developing RS generative foundation models. △ Less

Submitted 5 January, 2026; v1 submitted 29 December, 2025; originally announced December 2025.

FoldAct: Efficient and Stable Context Folding for Long-Horizon Search Agents

Authors: Jiaqi Shao, Yufeng Miao, Wei Zhang, Bing Luo

Abstract: Long-horizon reinforcement learning (RL) for large language models faces critical scalability challenges from unbounded context growth, leading to context folding methods that compress interaction history during task execution. However, existing approaches treat summary actions as standard actions, overlooking that summaries fundamentally modify the agent's future observation space, creating a pol… ▽ More Long-horizon reinforcement learning (RL) for large language models faces critical scalability challenges from unbounded context growth, leading to context folding methods that compress interaction history during task execution. However, existing approaches treat summary actions as standard actions, overlooking that summaries fundamentally modify the agent's future observation space, creating a policy-dependent, non-stationary observation distribution that violates core RL assumptions. This introduces three fundamental challenges: (1) gradient dilution where summary tokens receive insufficient training signal, (2) self-conditioning where policy updates change summary distributions, creating a vicious cycle of training collapse, and (3) computational cost from processing unique contexts at each turn. We introduce \textbf{FoldAct}\footnote{https://github.com/SHAO-Jiaqi757/FoldAct}, a framework that explicitly addresses these challenges through three key innovations: separated loss computation for independent gradient signals on summary and action tokens, full context consistency loss to reduce distribution shift, and selective segment training to reduce computational cost. Our method enables stable training of long-horizon search agents with context folding, addressing the non-stationary observation problem while improving training efficiency with 5.19$\times$ speedup. △ Less

Submitted 27 December, 2025; originally announced December 2025.

Automated stereotactic radiosurgery planning using a human-in-the-loop reasoning large language model agent

Authors: Humza Nusrat, Luke Francisco, Bing Luo, Hassan Bagher-Ebadian, Joshua Kim, Karen Chin-Snyder, Salim Siddiqui, Mira Shah, Eric Mellon, Mohammad Ghassemi, Anthony Doemer, Benjamin Movsas, Kundan Thind

Abstract: Stereotactic radiosurgery (SRS) demands precise dose shaping around critical structures, yet black-box AI systems have limited clinical adoption due to opacity concerns. We tested whether chain-of-thought reasoning improves agentic planning in a retrospective cohort of 41 patients with brain metastases treated with 18 Gy single-fraction SRS. We developed SAGE (Secure Agent for Generative Dose Expe… ▽ More Stereotactic radiosurgery (SRS) demands precise dose shaping around critical structures, yet black-box AI systems have limited clinical adoption due to opacity concerns. We tested whether chain-of-thought reasoning improves agentic planning in a retrospective cohort of 41 patients with brain metastases treated with 18 Gy single-fraction SRS. We developed SAGE (Secure Agent for Generative Dose Expertise), an LLM-based planning agent for automated SRS treatment planning. Two variants generated plans for each case: one using a non-reasoning model, one using a reasoning model. The reasoning variant showed comparable plan dosimetry relative to human planners on primary endpoints (PTV coverage, maximum dose, conformity index, gradient index; all p > 0.21) while reducing cochlear dose below human baselines (p = 0.022). When prompted to improve conformity, the reasoning model demonstrated systematic planning behaviors including prospective constraint verification (457 instances) and trade-off deliberation (609 instances), while the standard model exhibited none of these deliberative processes (0 and 7 instances, respectively). Content analysis revealed that constraint verification and causal explanation concentrated in the reasoning agent. The optimization traces serve as auditable logs, offering a path toward transparent automated planning. △ Less

Submitted 23 December, 2025; originally announced December 2025.

Vehicle-centric Perception via Multimodal Structured Pre-training

Authors: Wentao Wu, Xiao Wang, Chenglong Li, Jin Tang, Bin Luo

Abstract: Vehicle-centric perception plays a crucial role in many intelligent systems, including large-scale surveillance systems, intelligent transportation, and autonomous driving. Existing approaches lack effective learning of vehicle-related knowledge during pre-training, resulting in poor capability for modeling general vehicle perception representations. To handle this problem, we propose VehicleMAE-V… ▽ More Vehicle-centric perception plays a crucial role in many intelligent systems, including large-scale surveillance systems, intelligent transportation, and autonomous driving. Existing approaches lack effective learning of vehicle-related knowledge during pre-training, resulting in poor capability for modeling general vehicle perception representations. To handle this problem, we propose VehicleMAE-V2, a novel vehicle-centric pre-trained large model. By exploring and exploiting vehicle-related multimodal structured priors to guide the masked token reconstruction process, our approach can significantly enhance the model's capability to learn generalizable representations for vehicle-centric perception. Specifically, we design the Symmetry-guided Mask Module (SMM), Contour-guided Representation Module (CRM) and Semantics-guided Representation Module (SRM) to incorporate three kinds of structured priors into token reconstruction including symmetry, contour and semantics of vehicles respectively. SMM utilizes the vehicle symmetry constraints to avoid retaining symmetric patches and can thus select high-quality masked image patches and reduce information redundancy. CRM minimizes the probability distribution divergence between contour features and reconstructed features and can thus preserve holistic vehicle structure information during pixel-level reconstruction. SRM aligns image-text features through contrastive learning and cross-modal distillation to address the feature confusion caused by insufficient semantic understanding during masked reconstruction. To support the pre-training of VehicleMAE-V2, we construct Autobot4M, a large-scale dataset comprising approximately 4 million vehicle images and 12,693 text descriptions. Extensive experiments on five downstream tasks demonstrate the superior performance of VehicleMAE-V2. △ Less

Submitted 22 December, 2025; originally announced December 2025.

Comments: Journal extension of VehicleMAE (AAAI 2024)

MobileFineTuner: A Mobile-Native Framework for On-Device LLM Fine-Tuning in Real-World Embedded AI Applications

Authors: Jiaxiang Geng, Lunyu Zhao, Yiyi Lu, Bing Luo

Abstract: Large language models (LLMs) are moving from cloud-centric services toward on-device embedded AI, where models interact with private, longitudinal signals sensed from users and their physical environments. Mobile phones are a natural platform for such applications because they are continuously carried by users, connected to wearable sensors, and deeply integrated with daily mobile applications. Ho… ▽ More Large language models (LLMs) are moving from cloud-centric services toward on-device embedded AI, where models interact with private, longitudinal signals sensed from users and their physical environments. Mobile phones are a natural platform for such applications because they are continuously carried by users, connected to wearable sensors, and deeply integrated with daily mobile applications. However, practical LLM fine-tuning on commodity phones remains difficult. Existing fine-tuning frameworks are largely Python-based and server-oriented, making them hard to deploy inside mobile applications. We present MobileFineTuner, a mobile-native open-source framework for end-to-end LLM fine-tuning on commodity mobile phones. MobileFineTuner is implemented in C++ and provides a reusable training stack. To make fine-tuning feasible under mobile resource constraints, MobileFineTuner integrates a resource-aware training runtime with memory-efficient attention, activation checkpointing, gradient accumulation, parameter sharding, and energy-aware scheduling. We evaluate MobileFineTuner on real mobile phones using GPT-2, Gemma 3, and Qwen2.5 models across multiple fine-tuning tasks. The results show that MobileFineTuner reproduces standard Full-FT and LoRA fine-tuning behavior, substantially reduces memory pressure and improves executability on memory-constrained phones. We further demonstrate MobileFineTuner through a private campus health-agent application, where a local LLM is fine-tuned on user-specific wearable-sensing records to provide more personalized responses while keeping raw records on the phone. These results establish MobileFineTuner as a practical toolkit for studying and building on-device LLM fine-tuning applications in embedded AI and sensing systems. △ Less

Submitted 10 June, 2026; v1 submitted 8 December, 2025; originally announced December 2025.

Comments: 26 pages, 25 figures

Unleashing the Intrinsic Visual Representation Capability of Multimodal Large Language Models

Authors: Hengzhuang Li, Xinsong Zhang, Qiming Peng, Bin Luo, Han Hu, Dengyang Jiang, Han-Jia Ye, Teng Zhang, Hai Jin

Abstract: Multimodal Large Language Models (MLLMs) have demonstrated remarkable proficiency in multimodal tasks. Despite their impressive performance, MLLMs suffer from the modality imbalance issue, where visual information is often underutilized compared to textual representations in deeper layers, leading to degraded visual performance or hallucinations. This issue stems from the predominant reliance on n… ▽ More Multimodal Large Language Models (MLLMs) have demonstrated remarkable proficiency in multimodal tasks. Despite their impressive performance, MLLMs suffer from the modality imbalance issue, where visual information is often underutilized compared to textual representations in deeper layers, leading to degraded visual performance or hallucinations. This issue stems from the predominant reliance on next-text-token-prediction during training, which fails to provide direct visual supervisory signals, resulting in progressive homogenization of visual representations throughout the layers. To this end, we propose Latent Visual Reconstruction (LaVer), a novel training framework that facilitates MLLMs in learning more discriminative visual representations via masked image modeling in the joint latent semantic space of LLM. Our method offers direct visual activation to MLLMs, which exhibit increased visual attention allocation, indicating enhanced utilization of visual information. Extensive experiments across diverse benchmarks prove the superiority of our approach in various scenarios, especially those requiring dense visual capabilities. Code of LaVer is available at https://github.com/Fir-lat/LaVer. △ Less

Submitted 5 December, 2025; originally announced December 2025.

A Fine Evaluation Method for Cube Copying Test for Early Detection of Alzheimer's Disease

Authors: Xinyu Jiang, Cuiyun Gao, Wenda Huang, Yiyang Jiang, Binwen Luo, Yuxin Jiang, Mengting Wang, Haoran Wen, Yang Zhao, Xuemei Chen, Songqun Huang

Abstract: Background: Impairment of visual spatial cognitive function is the most common early clinical manifestation of Alzheimer's Disease (AD). When the Montreal Cognitive Assessment (MoCA) uses the "0/1" binary method ("pass/fail") to evaluate the visual spatial cognitive ability represented by the Cube Copying Test(CCT), the elder with less formal education generally score 0 point, resulting in serious… ▽ More Background: Impairment of visual spatial cognitive function is the most common early clinical manifestation of Alzheimer's Disease (AD). When the Montreal Cognitive Assessment (MoCA) uses the "0/1" binary method ("pass/fail") to evaluate the visual spatial cognitive ability represented by the Cube Copying Test(CCT), the elder with less formal education generally score 0 point, resulting in serious bias in the evaluation results. Therefore, this study proposes a fine evaluation method for CCT based on dynamic handwriting feature extraction of DH-SCSM-BLA. method : The Cogni-CareV3.0 software independently developed by our team was used to collect dynamic handwriting data of CCT. Then, the spatial and motion features of segmented dynamic handwriting were extracted, and feature matrix with unequal dimensions were normalized. Finally, a bidirectional long short-term memory network model combined with attention mechanism (BiLSTM-Attention) was adopted for classification. Result: The experimental results showed that: The proposed method has significant superiority compared to similar studies, with a classification accuracy of 86.69%. The distribution of cube drawing ability scores has significant regularity for three aspects such as MCI patients and healthy control group, age, and levels of education. It was also found that score for each cognitive task including cube drawing ability score is negatively correlated with age. Score for each cognitive task including cube drawing ability score, but positively correlated with levels of education significantly. Conclusion: This study provides a relatively objective and comprehensive evaluation method for early screening and personalized intervention of visual spatial cognitive impairment. △ Less

Submitted 1 December, 2025; originally announced December 2025.