arXiv subjects

Jun Zhao

Publications and source records attributed to Jun Zhao.

At least 91 records · Page 5Linked to original sources

arXiv2026

WideSeek: Advancing Wide Research via Multi-Agent Scaling

Search intelligence is evolving from Deep Research to Wide Research, a paradigm essential for retrieving and synthesizing comprehensive information under complex constraints in parallel. However, progress in this field is impeded by the lack of dedicated benchmarks and optimization methodologies for search breadth. To address these challenges, we take a deep dive into Wide Research from two perspectives: Data Pipeline and Agent Optimization. First, we produce WideSeekBench, a General Broad Information Seeking (GBIS) benchmark constructed via a rigorous multi-phase data pipeline to ensure diversity across the target information volume, logical constraints, and domains. Second, we introduce WideSeek, a dynamic hierarchical multi-agent architecture that can autonomously fork parallel sub-agents based on task requirements. Furthermore, we design a unified training framework that linearizes multi-agent trajectories and optimizes the system using end-to-end RL. Experimental results demonstrate the effectiveness of WideSeek and multi-agent RL, highlighting that scaling the number of agents is a promising direction for advancing the Wide Research paradigm.

cs.CL↗

arXiv2026

Safety, Mobility, and Environmental Impacts of Driver-Assistance-Enabled Electric Vehicles: An Empirical Study

The advancement of vehicle automation and the growing adoption of electric vehicles (EVs) are reshaping transportation systems. While fully automated vehicles are expected to improve traffic stability, efficiency, and sustainability, recent studies suggest that partially automated vehicles, such as those equipped with adaptive cruise control (ACC), may adversely affect traffic flow. These drawbacks may not extend to ACC-enabled EVs due to their distinct mechanical characteristics, including regenerative braking and smoother torque delivery. As a result, the impacts of EVs operating under ACC remain insufficiently understood. To address this gap, this study develops an empirical framework using the OpenACC dataset to compare ACC-enabled EVs and internal combustion engine vehicles. Dynamic time warping aligns comparable lead-vehicle trajectories. Results show that EVs exhibit smoother speed profiles, lower speed variability, and shorter spacing, leading to higher efficiency. EVs reduce critical safety events by over 85% and lower platoon-level emissions by up to 26.2%.

cs.ET↗

arXiv2026

Enhancing Image Quality Assessment Ability of LMMs via Retrieval-Augmented Generation

Large Multimodal Models (LMMs) have recently shown remarkable promise in low-level visual perception tasks, particularly in Image Quality Assessment (IQA), demonstrating strong zero-shot capability. However, achieving state-of-the-art performance often requires computationally expensive fine-tuning methods, which aim to align the distribution of quality-related token in output with image quality levels. Inspired by recent training-free works for LMM, we introduce IQARAG, a novel, training-free framework that enhances LMMs' IQA ability. IQARAG leverages Retrieval-Augmented Generation (RAG) to retrieve some semantically similar but quality-variant reference images with corresponding Mean Opinion Scores (MOSs) for input image. These retrieved images and input image are integrated into a specific prompt. Retrieved images provide the LMM with a visual perception anchor for IQA task. IQARAG contains three key phases: Retrieval Feature Extraction, Image Retrieval, and Integration & Quality Score Generation. Extensive experiments across multiple diverse IQA datasets, including KADID, KonIQ, LIVE Challenge, and SPAQ, demonstrate that the proposed IQARAG effectively boosts the IQA performance of LMMs, offering a resource-efficient alternative to fine-tuning for quality assessment.

cs.CV↗

arXiv2026

Interactive visualizations for adolescents to understand and challenge algorithmic profiling in online platforms

Social media platforms regularly track, aggregate, and monetize adolescents' data, yet provide them with little visibility or agency over how algorithms construct their digital identities and make inferences about them. We introduce Algorithmic Mirror, an interactive visualization tool that transforms opaque profiling practices into explorable landscapes of personal data. It uniquely leverages adolescents' real digital footprints across YouTube, TikTok, and Netflix, to provide situated, personalized insights into datafication over time. In our study with 27 participants (ages 12--16), we show how engaging with their own data enabled adolescents to uncover the scale and persistence of data collection, recognize cross-platform profiling, and critically reflect algorithmic categorizations of their interests. These findings highlight how identity is a powerful motivator for adolescents' desire for greater digital agency, underscoring the need for platforms and policymakers to move toward structural reforms that guarantee children better transparency and the agency to influence their online experiences.

cs.HC↗

arXiv2026

Learning How to Remember: A Meta-Cognitive Management Method for Structured and Transferable Agent Memory

Large language model (LLM) agents increasingly rely on accumulated memory to solve long-horizon decision-making tasks. However, most existing approaches store memory in fixed representations and reuse it at a single or implicit level of abstraction, which limits generalization and often leads to negative transfer when distribution shift. This paper proposes the Meta-Cognitive Memory Abstraction method (MCMA), which treats memory abstraction as a learnable cognitive skill rather than a fixed design choice. MCMA decouples task execution from memory management by combining a frozen task model with a learned memory copilot. The memory copilot is trained using direct preference optimization, it determines how memories should be structured, abstracted, and reused. Memories are further organized into a hierarchy of abstraction levels, enabling selective reuse based on task similarity. When no memory is transferable, MCMA transfers the ability to abstract and manage memory by transferring the memory copilot. Experiments on ALFWorld, ScienceWorld, and BabyAI demonstrate substantial improvements in performance, out-of-distribution generalization, and cross-task transfer over several baselines.

cs.AI↗

arXiv2026

Kitaev interactions in the van der Waals antiferromagnet VBr3

Van der Waals materials hosting Kitaev interactions are promising platforms for exploring exotic quantum phenomena. Here, we report inelastic neutron scattering investigations of the van der Waals antiferromagnet VBr3, which forms a honeycomb lattice structure at room temperature and exhibits zigzag-type magnetic order below 26.5 K. Our observations reveal distinctive low-energy spin excitations arising from gamma, gamma', and M' points, each featuring a spin gap of around 2.5 meV. The overall spin excitation spectra can be effectively described by a spin Hamiltonian incorporating substantial nearest-neighbor Kitaev and biquadratic interactions, along with Heisenberg interactions. Our findings not only establish VBr3 as a new Kitaev magnet but also suggest that ligand engineering may provide a promising strategy to modulate Kitaev interactions, offering new opportunities for designing Kitaev materials with tailored quantum properties.

cond-mat.str-el↗

arXiv2026

Parameter Training Efficiency Aware Resource Allocation for AIGC in Space-Air-Ground Integrated Networks

With the evolution of artificial intelligence-generated content (AIGC) techniques and the development of space-air-ground integrated networks (SAGIN), there will be a growing opportunity to enhance more users' mobile experience with customized AIGC applications. This is made possible through the use of parameter-efficient fine-tuning (PEFT) training alongside mobile edge computing. In this paper, we formulate the optimization problem of maximizing the parameter training efficiency of the SAGIN system over wireless networks under limited resource constraints. We propose the Parameter training efficiency Aware Resource Allocation (PARA) technique to jointly optimize user association, data offloading, and communication and computational resource allocation. Solid proofs are presented to solve this difficult sum of ratios problem based on quadratically constrained quadratic programming (QCQP), semidefinite programming (SDP), graph theory, and fractional programming (FP) techniques. Our proposed PARA technique is effective in finding a stationary point of this non-convex problem. The simulation results demonstrate that the proposed PARA method outperforms other baselines.

cs.ET↗

arXiv2025

DE$^3$-BERT: Distance-Enhanced Early Exiting for BERT based on Prototypical Networks

Early exiting has demonstrated its effectiveness in accelerating the inference of pre-trained language models like BERT by dynamically adjusting the number of layers executed. However, most existing early exiting methods only consider local information from an individual test sample to determine their exiting indicators, failing to leverage the global information offered by sample population. This leads to suboptimal estimation of prediction correctness, resulting in erroneous exiting decisions. To bridge the gap, we explore the necessity of effectively combining both local and global information to ensure reliable early exiting during inference. Purposefully, we leverage prototypical networks to learn class prototypes and devise a distance metric between samples and class prototypes. This enables us to utilize global information for estimating the correctness of early predictions. On this basis, we propose a novel Distance-Enhanced Early Exiting framework for BERT (DE$^3$-BERT). DE$^3$-BERT implements a hybrid exiting strategy that supplements classic entropy-based local information with distance-based global information to enhance the estimation of prediction correctness for more reliable early exiting decisions. Extensive experiments on the GLUE benchmark demonstrate that DE$^3$-BERT consistently outperforms state-of-the-art models under different speed-up ratios with minimal storage or computational overhead, yielding a better trade-off between model performance and inference efficiency. Additionally, an in-depth analysis further validates the generality and interpretability of our method.

cs.LG↗

arXiv2025

Reasoning or Memorization? Unreliable Results of Reinforcement Learning Due to Data Contamination

Reasoning in large language models has long been a central research focus, and recent studies employing reinforcement learning (RL) have introduced diverse methods that yield substantial performance gains with minimal or even no external supervision. Surprisingly, some studies even suggest that random or incorrect reward signals can enhance performance. However, these breakthroughs are predominantly observed for the mathematically strong Qwen2.5 series on benchmarks such as MATH-500, AMC, and AIME, and seldom transfer to models like Llama, which warrants a more in-depth investigation. In this work, our empirical analysis reveals that pre-training on massive web-scale corpora leaves Qwen2.5 susceptible to data contamination in widely used benchmarks. Consequently, conclusions derived from contaminated benchmarks on Qwen2.5 series may be unreliable. To obtain trustworthy evaluation results, we introduce a generator that creates fully clean arithmetic problems of arbitrary length and difficulty, dubbed RandomCalculation. Using this leakage-free dataset, we show that only accurate reward signals yield steady improvements that surpass the base model's performance boundary in mathematical reasoning, whereas random or incorrect rewards do not. Moreover, we conduct more fine-grained analyses to elucidate the factors underlying the different performance observed on the MATH-500 and RandomCalculation benchmarks. Consequently, we recommend that future studies evaluate models on uncontaminated benchmarks and, when feasible, test various model series to ensure trustworthy conclusions about RL and related methods.

cs.LG↗

arXiv2025

Game-RL: Synthesizing Multimodal Verifiable Game Data to Boost VLMs' General Reasoning

Vision-language reinforcement learning (RL) has primarily focused on narrow domains (e.g. geometry or chart reasoning). This leaves broader training scenarios and resources underexplored, limiting the exploration and learning of Vision Language Models (VLMs) through RL. We find video games inherently provide rich visual elements and mechanics that are easy to verify. To fully use the multimodal and verifiable reward in video games, we propose Game-RL, constructing diverse game tasks for RL training to boost VLMs general reasoning ability. To obtain training data, we propose Code2Logic, a novel approach that adapts game code to synthesize game reasoning task data, thus obtaining the GameQA dataset of 30 games and 158 tasks with controllable difficulty gradation. Unexpectedly, RL training solely on GameQA enables multiple VLMs to achieve performance improvements across 7 diverse vision-language benchmarks, demonstrating the value of Game-RL for enhancing VLMs' general reasoning. Furthermore, this suggests that video games may serve as valuable scenarios and resources to boost general reasoning abilities. Our code, dataset and models are available at the GitHub repository.

cs.CL↗

arXiv2025

EvoEdit: Lifelong Free-Text Knowledge Editing through Latent Perturbation Augmentation and Knowledge-driven Parameter Fusion

Adjusting the outdated knowledge of large language models (LLMs) after deployment remains a major challenge. This difficulty has spurred the development of knowledge editing, which seeks to accurately and efficiently modify a model's internal (parametric) knowledge without retraining it from scratch. However, existing methods suffer from two limitations. First, they depend on structured triplets that are misaligned with the free-text nature of LLM pretraining and fail to capture the nuanced relationships among facts. Second, they typically support one-time knowledge updates, with relatively limited research on the problem of sequential or lifelong editing. To address these gaps, we propose a new task, Lifelong Free-text Knowledge Editing (LF-Edit), which enables models to incorporate updates expressed in natural language and supports continual editing over time. Despite its promise, LF-Edit faces the dual challenge of integrating new knowledge while mitigating the forgetting of prior information. To foster research on this new task, we construct a large-scale benchmark, Multi-Rank Lifelong Free-text Editing Benchmark (MRLF-Bench), containing 16,835 free-text edit requests. We further design a cognitively inspired multi-rank evaluation framework encompassing four levels: memorization, understanding, constrained comprehension, and reasoning. To tackle the challenges inherent in LF-Edit, we introduce a novel approach named EvoEdit that enhances knowledge injection through Latent Perturbation Augmentation and preserves prior information via Knowledge-driven Parameter Fusion. Experimental results demonstrate that EvoEdit substantially outperforms existing knowledge editing methods on the proposed LF-Edit task.

cs.CL↗

arXiv2025

DAComp: Benchmarking Data Agents across the Full Data Intelligence Lifecycle

Real-world enterprise data intelligence workflows encompass data engineering that turns raw sources into analytical-ready tables and data analysis that convert those tables into decision-oriented insights. We introduce DAComp, a benchmark of 210 tasks that mirrors these complex workflows. Data engineering (DE) tasks require repository-level engineering on industrial schemas, including designing and building multi-stage SQL pipelines from scratch and evolving existing systems under evolving requirements. Data analysis (DA) tasks pose open-ended business problems that demand strategic planning, exploratory analysis through iterative coding, interpretation of intermediate results, and the synthesis of actionable recommendations. Engineering tasks are scored through execution-based, multi-metric evaluation. Open-ended tasks are assessed by a reliable, experimentally validated LLM-judge, which is guided by hierarchical, meticulously crafted rubrics. Our experiments reveal that even state-of-the-art agents falter on DAComp. Performance on DE tasks is particularly low, with success rates under 20%, exposing a critical bottleneck in holistic pipeline orchestration, not merely code generation. Scores on DA tasks also average below 40%, highlighting profound deficiencies in open-ended reasoning and demonstrating that engineering and analysis are distinct capabilities. By clearly diagnosing these limitations, DAComp provides a rigorous and realistic testbed to drive the development of truly capable autonomous data agents for enterprise settings. Our data and code are available at https://da-comp.github.io

cs.CL↗

arXiv2025

A Secure Semantic Communication System Based on Knowledge Graph

This study proposes a novel approach to ensure the security of textual data transmission in a semantic communication system. In the proposed system, a sender transmits textual information to a receiver, while a potential eavesdropper attempts to intercept the information. At the sender side, the text is initially preprocessed, where each sentence is annotated with its corresponding topic, and subsequently extracted into a knowledge graph. To achieve the secure transmission of the knowledge graph, we propose a channel encryption scheme that integrates constellation diagonal transformation with multi-parameter weighted fractional Fourier transform (MP-WFRFT). At the receiver side, the textual data is first decrypted, and then recovered via a transformer model. Experimental results demonstrate that the proposed method reduces the probability of information compromise. The legitimate receiver achieves a Bilingual Evaluation Understudy (BLEU) score of 0.9, whereas the BLEU score of the eavesdropper remains below 0.3. Compared to the baselines, the proposed method can improve the security by up to 20%.

cs.CR↗

arXiv2025

Bias-Restrained Prefix Representation Finetuning for Mathematical Reasoning

Parameter-Efficient finetuning (PEFT) enhances model performance on downstream tasks by updating a minimal subset of parameters. Representation finetuning (ReFT) methods further improve efficiency by freezing model weights and optimizing internal representations with fewer parameters than PEFT, outperforming PEFT on several tasks. However, ReFT exhibits a significant performance decline on mathematical reasoning tasks. To address this problem, the paper demonstrates that ReFT's poor performance on mathematical tasks primarily stems from its struggle to generate effective reasoning prefixes during the early inference phase. Moreover, ReFT disturbs the numerical encoding and the error accumulats during the CoT stage. Based on these observations, this paper proposes Bias-REstrained Prefix Representation FineTuning (BREP ReFT), which enhances ReFT's mathematical reasoning capability by truncating training data to optimize the generation of initial reasoning prefixes, intervening on the early inference stage to prevent error accumulation, and constraining the intervention vectors' magnitude to avoid disturbing numerical encoding. Extensive experiments across diverse model architectures demonstrate BREP's superior effectiveness, efficiency, and robust generalization capability, outperforming both standard ReFT and weight-based PEFT methods on the task of mathematical reasoning. The source code is available at https://github.com/LiangThree/BREP.

cs.LG↗

arXiv2025

SparK: Query-Aware Unstructured Sparsity with Recoverable KV Cache Channel Pruning

Long-context inference in large language models (LLMs) is increasingly constrained by the KV cache bottleneck: memory usage grows linearly with sequence length, while attention computation scales quadratically. Existing approaches address this issue by compressing the KV cache along the temporal axis through strategies such as token eviction or merging to reduce memory and computational overhead. However, these methods often neglect fine-grained importance variations across feature dimensions (i.e., the channel axis), thereby limiting their ability to effectively balance efficiency and model accuracy. In reality, we observe that channel saliency varies dramatically across both queries and positions: certain feature channels carry near-zero information for a given query, while others spike in relevance. To address this oversight, we propose SPARK, a training-free plug-and-play method that applies unstructured sparsity by pruning KV at the channel level, while dynamically restoring the pruned entries during attention score computation. Notably, our approach is orthogonal to existing KV compression and quantization techniques, making it compatible for integration with them to achieve further acceleration. By reducing channel-level redundancy, SPARK enables processing of longer sequences within the same memory budget. For sequences of equal length, SPARK not only preserves or improves model accuracy but also reduces KV cache storage by over 30% compared to eviction-based methods. Furthermore, even with an aggressive pruning ratio of 80%, SPARK maintains performance with less degradation than 5% compared to the baseline eviction method, demonstrating its robustness and effectiveness. Our code will be available at https://github.com/Xnhyacinth/SparK.

cs.CL↗

arXiv2025

Hardware-aligned Hierarchical Sparse Attention for Efficient Long-term Memory Access

A key advantage of Recurrent Neural Networks (RNNs) over Transformers is their linear computational and space complexity enables faster training and inference for long sequences. However, RNNs are fundamentally unable to randomly access historical context, and simply integrating attention mechanisms may undermine their efficiency advantages. To overcome this limitation, we propose Hierarchical Sparse Attention (HSA), a novel attention mechanism that enhances RNNs with long-range random access flexibility while preserving their merits in efficiency and length generalization. HSA divides inputs into chunks, selects the top-$k$ chunks and hierarchically aggregates information. The core innovation lies in learning token-to-chunk relevance based on fine-grained token-level information inside each chunk. This approach enhances the precision of chunk selection across both in-domain and out-of-domain context lengths. To make HSA efficient, we further introduce a hardware-aligned kernel design. By combining HSA with Mamba, we introduce RAMba, which achieves perfect accuracy in passkey retrieval across 64 million contexts despite pre-training on only 4K-length contexts, and significant improvements on various downstream tasks, with nearly constant memory footprint. These results show RAMba's huge potential in long-context modeling.

cs.CL↗

arXiv2025

InfoFlow: Reinforcing Search Agent Via Reward Density Optimization

Reinforcement Learning with Verifiable Rewards (RLVR) is a promising approach for enhancing agentic deep search. However, its application is often hindered by low \textbf{Reward Density} in deep search scenarios, where agents expend significant exploratory costs for infrequent and often null final rewards. In this paper, we formalize this challenge as the \textbf{Reward Density Optimization} problem, which aims to improve the reward obtained per unit of exploration cost. This paper introduce \textbf{InfoFlow}, a systematic framework that tackles this problem from three aspects. 1) \textbf{Subproblem decomposition}: breaking down long-range tasks to assign process rewards, thereby providing denser learning signals. 2) \textbf{Failure-guided hints}: injecting corrective guidance into stalled trajectories to increase the probability of successful outcomes. 3) \textbf{Dual-agent refinement}: employing a dual-agent architecture to offload the cognitive burden of deep exploration. A refiner agent synthesizes the search history, which effectively compresses the researcher's perceived trajectory, thereby reducing exploration cost and increasing the overall reward density. We evaluate InfoFlow on multiple agentic search benchmarks, where it significantly outperforms strong baselines, enabling lightweight LLMs to achieve performance comparable to advanced proprietary LLMs.

cs.CL↗

arXiv2025

Omni-Reward: Towards Generalist Omni-Modal Reward Modeling with Free-Form Preferences

Reward models (RMs) play a critical role in aligning AI behaviors with human preferences, yet they face two fundamental challenges: (1) Modality Imbalance, where most RMs are mainly focused on text and image modalities, offering limited support for video, audio, and other modalities; and (2) Preference Rigidity, where training on fixed binary preference pairs fails to capture the complexity and diversity of personalized preferences. To address the above challenges, we propose Omni-Reward, a step toward generalist omni-modal reward modeling with support for free-form preferences, consisting of: (1) Evaluation: We introduce Omni-RewardBench, the first omni-modal RM benchmark with free-form preferences, covering nine tasks across five modalities including text, image, video, audio, and 3D; (2) Data: We construct Omni-RewardData, a multimodal preference dataset comprising 248K general preference pairs and 69K instruction-tuning pairs for training generalist omni-modal RMs; (3) Model: We propose Omni-RewardModel, which includes both discriminative and generative RMs, and achieves strong performance on Omni-RewardBench as well as other widely used reward modeling benchmarks.

cs.CL↗