arXiv subjects

Jinjie Gu

Publications and source records attributed to Jinjie Gu.

At least 37 records · Page 2Linked to original sources

arXiv2026

GroupRank: A Groupwise Paradigm for Effective and Efficient Passage Reranking with LLMs

Large Language Models (LLMs) have emerged as powerful tools for passage reranking in information retrieval, leveraging their superior reasoning capabilities to address the limitations of conventional models on complex queries. However, current LLM-based reranking paradigms are fundamentally constrained by an efficiency-accuracy trade-off: (1) pointwise methods are efficient but ignore inter-document comparison, yielding suboptimal accuracy; (2) listwise methods capture global context but suffer from context-window constraints and prohibitive inference latency. To address these issues, we propose GroupRank, a novel paradigm that balances flexibility and context awareness. To unlock the full potential of groupwise reranking, we propose an answer-free data synthesis pipeline that fuses local pointwise signals with global listwise rankings. These samples facilitate supervised fine-tuning and reinforcement learning, with the latter guided by a specialized group-ranking reward comprising ranking-utility and group-alignment. These complementary components synergistically optimize document ordering and score calibration to reflect intrinsic query-document relevance. Experimental results show GroupRank achieves a state-of-the-art 65.2 NDCG@10 on BRIGHT and surpasses baselines by 2.1 points on R2MED, while delivering a 6.4$\times$ inference speedup.

cs.IR↗

arXiv2026

DIVER: A Multi-Stage Approach for Reasoning-intensive Information Retrieval

Retrieval-augmented generation has achieved strong performance on knowledge-intensive tasks where query-document relevance can be identified through direct lexical or semantic matches. However, many real-world queries involve abstract reasoning, analogical thinking, or multi-step inference, which existing retrievers often struggle to capture. To address this challenge, we present DIVER, a retrieval pipeline designed for reasoning-intensive information retrieval. It consists of four components. The document preprocessing stage enhances readability and preserves content by cleaning noisy texts and segmenting long documents. The query expansion stage leverages large language models to iteratively refine user queries with explicit reasoning and evidence from retrieved documents. The retrieval stage employs a model fine-tuned on synthetic data spanning medical and mathematical domains, along with hard negatives, enabling effective handling of reasoning-intensive queries. Finally, the reranking stage combines pointwise and listwise strategies to produce both fine-grained and globally consistent rankings. On the BRIGHT benchmark, DIVER achieves state-of-the-art nDCG@10 scores of 46.8 overall and 31.9 on original queries, consistently outperforming competitive reasoning-aware models. These results demonstrate the effectiveness of reasoning-aware retrieval strategies in complex real-world tasks.

cs.IR↗

arXiv2026

A Tri-Modal Dataset and a Baseline System for Tracking Unmanned Aerial Vehicles

With the proliferation of low altitude unmanned aerial vehicles (UAVs), visual multi-object tracking is becoming a critical security technology, demanding significant robustness even in complex environmental conditions. However, tracking UAVs using a single visual modality often fails in challenging scenarios, such as low illumination, cluttered backgrounds, and rapid motion. Although multi-modal multi-object UAV tracking is more resilient, the development of effective solutions has been hindered by the absence of dedicated public datasets. To bridge this gap, we release MM-UAV, the first large-scale benchmark for Multi-Modal UAV Tracking, integrating three key sensing modalities, e.g. RGB, infrared (IR), and event signals. The dataset spans over 30 challenging scenarios, with 1,321 synchronised multi-modal sequences, and more than 2.8 million annotated frames. Accompanying the dataset, we provide a novel multi-modal multi-UAV tracking framework, designed specifically for UAV tracking applications and serving as a baseline for future research. Our framework incorporates two key technical innovations, e.g. an offset-guided adaptive alignment module to resolve spatio mismatches across sensors, and an adaptive dynamic fusion module to balance complementary information conveyed by different modalities. Furthermore, to overcome the limitations of conventional appearance modelling in multi-object tracking, we introduce an event-enhanced association mechanism that leverages motion cues from the event modality for more reliable identity maintenance. Comprehensive experiments demonstrate that the proposed framework consistently outperforms state-of-the-art methods. To foster further research in multi-modal UAV tracking, both the dataset and source code will be made publicly available at https://xuefeng-zhu5.github.io/MM-UAV/.

cs.CV↗

arXiv2026

Collaborating Vision, Depth, and Thermal Signals for Multi-Modal Tracking: Dataset and Algorithm

Existing multi-modal object tracking approaches primarily focus on dual-modal paradigms, such as RGB-Depth or RGB-Thermal, yet remain challenged in complex scenarios due to limited input modalities. To address this gap, this work introduces a novel multi-modal tracking task that leverages three complementary modalities, including visible RGB, Depth (D), and Thermal Infrared (TIR), aiming to enhance robustness in complex scenarios. To support this task, we construct a new multi-modal tracking dataset, coined RGBDT500, which consists of 500 videos with synchronised frames across the three modalities. Each frame provides spatially aligned RGB, depth, and thermal infrared images with precise object bounding box annotations. Furthermore, we propose a novel multi-modal tracker, dubbed RDTTrack. RDTTrack integrates tri-modal information for robust tracking by leveraging a pretrained RGB-only tracking model and prompt learning techniques. In specific, RDTTrack fuses thermal infrared and depth modalities under a proposed orthogonal projection constraint, then integrates them with RGB signals as prompts for the pre-trained foundation tracking model, effectively harmonising tri-modal complementary cues. The experimental results demonstrate the effectiveness and advantages of the proposed method, showing significant improvements over existing dual-modal approaches in terms of tracking accuracy and robustness in complex scenarios. The dataset and source code are publicly available at https://xuefeng-zhu5.github.io/RGBDT500.

cs.CV↗

arXiv2026

LiveAgentBench: Comprehensive Benchmarking of Agentic Systems Across 104 Real-World Challenges

As large language models grow more capable, general AI agents have become increasingly prevalent in practical applications. However, existing benchmarks face significant limitations, failing to represent real-world user tasks accurately. To address this gap, we present LiveAgentBench, a comprehensive benchmark with 104 scenarios that reflect real user requirements. It is constructed from publicly sourced questions on social media and real-world products. Central to our approach is the Social Perception-Driven Data Generation (SPDG) method, a novel process we developed to ensure each question's real-world relevance, task complexity, and result verifiability. We evaluate various models, frameworks, and commercial products using LiveAgentBench, revealing their practical performance and identifying areas for improvement. This release includes 374 tasks, with 125 for validation and 249 for testing. The SPDG process enables continuous updates with fresh queries from real-world interactions.

cs.AI↗

arXiv2026

LiveClin: A Live Clinical Benchmark without Leakage

The reliability of medical LLM evaluation is critically undermined by data contamination and knowledge obsolescence, leading to inflated scores on static benchmarks. To address these challenges, we introduce LiveClin, a live benchmark designed for approximating real-world clinical practice. Built from contemporary, peer-reviewed case reports and updated biannually, LiveClin ensures clinical currency and resists data contamination. Using a verified AI-human workflow involving 239 physicians, we transform authentic patient cases into complex, multimodal evaluation scenarios that span the entire clinical pathway. The benchmark currently comprises 1,407 case reports and 6,605 questions. Our evaluation of 26 models on LiveClin reveals the profound difficulty of these real-world scenarios, with the top-performing model achieving a Case Accuracy of just 35.7%. In benchmarking against human experts, Chief Physicians achieved the highest accuracy, followed closely by Attending Physicians, with both surpassing most models. LiveClin thus provides a continuously evolving, clinically grounded framework to guide the development of medical LLMs towards closing this gap and achieving greater reliability and real-world utility. Our data and code are publicly available at https://github.com/AQ-MedAI/LiveClin.

cs.LG↗

arXiv2026

ClinAlign: Scaling Healthcare Alignment from Clinician Preference

Although large language models (LLMs) demonstrate expert-level medical knowledge, aligning their open-ended outputs with fine-grained clinician preferences remains challenging. Existing methods often rely on coarse objectives or unreliable automated judges that are weakly grounded in professional guidelines. We propose a two-stage framework to address this gap. First, we introduce HealthRubrics, a dataset of 7,034 physician-verified preference examples in which clinicians refine LLM-drafted rubrics to meet rigorous medical standards. Second, we distill these rubrics into HealthPrinciples: 119 broadly reusable, clinically grounded principles organized by clinical dimensions, enabling scalable supervision beyond manual annotation. We use HealthPrinciples for (1) offline alignment by synthesizing rubrics for unlabeled queries and (2) an inference-time tool for guided self-revision. A 30B-A3B model trained with our framework achieves 33.4% on HealthBench-Hard, outperforming much larger models including Deepseek-R1 and o3, establishing a resource-efficient baseline for clinical alignment.

cs.AI↗

arXiv2026

Evolving Interactive Diagnostic Agents in a Virtual Clinical Environment

We present a framework for training large language models (LLMs) as diagnostic agents with reinforcement learning, enabling them to manage multi-turn interactive diagnostic processes, adaptively select examinations, and commit to final diagnoses. Unlike instruction-tuned models trained on static data, our method acquires diagnostic strategies through dynamic exploration and outcome-based feedback, mapping evolving patient states to the next optimal examination and subsequent diagnosis. Our contributions include: (i) DiagGym, a diagnostics world model trained with electronic health records, serving as a virtual clinical environment to support closed-loop in-silico training and evaluation for interactive diagnosis; (ii) DiagAgent, trained via end-to-end multi-turn RL to learn dynamic diagnostic policies that optimize both interactive effectiveness and final accuracy; (iii) DiagBench, a multi-center diagnostic benchmark designed to evaluate multi-turn diagnostic interaction trajectories. The benchmark comprises 2.2K physician-validated cases sourced from 4 distinct distributions, alongside 3.3K physician-written rubrics for granular process-oriented evaluation. (iv) Extensive evaluations demonstrate DiagAgent's superior performance across both in-domain and out-of-domain (OOD) settings. DiagAgent significantly outperforms 11 SOTA LLMs and 2 prompt-engineered agents. In the end-to-end setting, it delivers a 11.20% increase in diagnostic accuracy and a 17.58% boost in examination recommendation F1 score, while consistently maintaining SOTA performance across all three external centers. Furthermore, in rubric-based evaluations, it surpasses the next-best model by 7.1% in weighted rubric score. These findings indicate that learning policies in interactive clinical environments confers long-term diagnostic management abilities unattainable through passive training.

cs.CL↗

arXiv2026

CTRL-RAG: Contrastive Likelihood Reward Based Reinforcement Learning for Context-Faithful RAG Models

With the growing use of Retrieval-Augmented Generation (RAG), training large language models (LLMs) for context-sensitive reasoning and faithfulness is increasingly important. Existing RAG-oriented reinforcement learning (RL) methods rely on external rewards that often fail to evaluate document faithfulness, and may misjudge similar answers in open-domain settings. In addition, there is no RAG-based selfreward mechanism. Moreover, although such a mechanism could in principle estimate answer confidence given documents, the absence of objective feedback in a self-judgment can cause hallucination accumulation and eventual model collapse. To tackle these issues, we propose a novel "internal-external" hybrid reward framework centered on a Contrastive Likelihood Reward (CLR). CLR directly optimizes the log-likelihood gap between responses conditioned on prompts with and without supporting evidence. This encourages the model to extract relevant evidence and increases its confidence when grounded in a specific context. Experiments show that our method (used alone or combined with external correctness rewards) achieves strong performance on singlehop, multi-hop, vertical-domain, and faithfulness benchmarks. Our training code and models are coming soon.

cs.CL↗

arXiv2026

Don't Just Fine-tune the Agent, Tune the Environment

Large Language Model (LLM) agents show great promise for complex, multi-turn tool-use tasks, but their development is often hampered by the extreme scarcity of high-quality training data. Supervised fine-tuning (SFT) on synthetic data leads to overfitting, whereas standard reinforcement learning (RL) struggles with a critical cold-start problem and training instability. To address these challenges, we introduce $\textbf{Environment Tuning}$, a novel training paradigm that enables agents to learn complex behaviors directly from problem instances without relying on pre-collected expert trajectories. $\textbf{Environment Tuning}$ orchestrates this learning process through a structured curriculum, actionable environment augmentation that provides corrective feedback, and fine-grained progress rewards to ensure stable and efficient exploration. Using only 400 problem instances from Berkeley Function-Calling Leaderboard (BFCL) benchmark, our method not only achieves competitive in-distribution performance against strong baselines but also demonstrates superior out-of-distribution generalization, overcoming the performance collapse common to SFT-based approaches. Our work presents a paradigm shift from supervised fine-tuning on static trajectories to dynamic, environment-based exploration, paving the way for training more robust and data-efficient agents. The code is available at https://github.com/inclusionAI/AWorld-RL/tree/main/EnvTuning.

cs.AI↗

arXiv2026

V2P: Visual Attention Calibration for GUI Grounding via Background Suppression and Center Peaking

Precise localization of GUI elements is crucial for the development of GUI agents. Traditional methods rely on bounding box or center-point regression, neglecting spatial interaction uncertainty and visual-semantic hierarchies. Recent methods incorporate attention mechanisms but still face two key issues: (1) ignoring processing background regions causes attention drift from the desired area, and (2) uniform modeling the target UI element fails to distinguish between its center and edges, leading to click imprecision. Inspired by how humans visually process and interact with GUI elements, we propose the Valley-to-Peak (V2P) method to address these issues. To mitigate background distractions, V2P introduces a suppression attention mechanism that minimizes the model's focus on irrelevant regions to highlight the intended region. For the issue of center-edge distinction, V2P applies a Fitts' Law-inspired approach by modeling GUI interactions as 2D Gaussian heatmaps where the weight gradually decreases from the center towards the edges. The weight distribution follows a Gaussian function, with the variance determined by the target's size. Consequently, V2P effectively isolates the target area and teaches the model to concentrate on the most essential point of the UI element. The model trained by V2P achieves the performance with 92.4\% and 52.5\% on two benchmarks ScreenSpot-v2 and ScreenSpot-Pro. Ablations further confirm each component's contribution, underscoring V2P's generalizability in precise GUI grounding tasks and its potential for real-world deployment in future GUI agents.

cs.AI↗

arXiv2026

V2P: Visual Attention Calibration for GUI Grounding via Background Suppression and Center Peaking

Precise localization of GUI elements is crucial for the development of GUI agents. Traditional methods rely on bounding box or center-point regression, neglecting spatial interaction uncertainty and visual-semantic hierarchies. Recent methods incorporate attention mechanisms but still face two key issues: (1) ignoring processing background regions causes attention drift from the desired area, and (2) uniform modeling the target UI element fails to distinguish between its center and edges, leading to click imprecision. Inspired by how humans visually process and interact with GUI elements, we propose the Valley-to-Peak (V2P) method to address these issues. To mitigate background distractions, V2P introduces a suppression attention mechanism that minimizes the model's focus on irrelevant regions to highlight the intended region. For the issue of center-edge distinction, V2P applies a Fitts' Law-inspired approach by modeling GUI interactions as 2D Gaussian heatmaps where the weight gradually decreases from the center towards the edges. The weight distribution follows a Gaussian function, with the variance determined by the target's size. Consequently, V2P effectively isolates the target area and teaches the model to concentrate on the most essential point of the UI element. The model trained by V2P achieves the performance with 92.4\% and 52.5\% on two benchmarks ScreenSpot-v2 and ScreenSpot-Pro (see Fig.~\ref{fig:main_results_charts}). Ablations further confirm each component's contribution, underscoring V2P's generalizability in precise GUI grounding tasks and its potential for real-world deployment in future GUI agents.

cs.AI↗

arXiv2026

RAG-R1: Incentivizing the Search and Reasoning Capabilities of LLMs through Multi-query Parallelism

Large Language Models (LLMs), despite their remarkable capabilities, are prone to generating hallucinated or outdated content due to their static internal knowledge. While Retrieval-Augmented Generation (RAG) integrated with Reinforcement Learning (RL) offers a solution, these methods are fundamentally constrained by a single-query mode, leading to prohibitive latency and inherent brittleness. To overcome these limitations, we introduce RAG-R1, a novel two-stage training framework centered around multi-query parallelism. Our framework enables LLMs to adaptively leverage internal and external knowledge during the reasoning process while transitioning from the single-query mode to multi-query parallelism. This architectural shift bolsters reasoning robustness while significantly reducing inference latency. Extensive experiments on seven question-answering benchmarks confirm the superiority of our method, which outperforms the strongest baseline by up to 13.7% and decreases inference time by 11.1%.

cs.CL↗

arXiv2026

MLB: A Scenario-Driven Benchmark for Evaluating Large Language Models in Clinical Applications

The proliferation of Large Language Models (LLMs) presents transformative potential for healthcare, yet practical deployment is hindered by the absence of frameworks that assess real-world clinical utility. Existing benchmarks test static knowledge, failing to capture the dynamic, application-oriented capabilities required in clinical practice. To bridge this gap, we introduce a Medical LLM Benchmark MLB, a comprehensive benchmark evaluating LLMs on both foundational knowledge and scenario-based reasoning. MLB is structured around five core dimensions: Medical Knowledge (MedKQA), Safety and Ethics (MedSE), Medical Record Understanding (MedRU), Smart Services (SmartServ), and Smart Healthcare (SmartCare). The benchmark integrates 22 datasets (17 newly curated) from diverse Chinese clinical sources, covering 64 clinical specialties. Its design features a rigorous curation pipeline involving 300 licensed physicians. Besides, we provide a scalable evaluation methodology, centered on a specialized judge model trained via Supervised Fine-Tuning (SFT) on expert annotations. Our comprehensive evaluation of 10 leading models reveals a critical translational gap: while the top-ranked model, Kimi-K2-Instruct (77.3% accuracy overall), excels in structured tasks like information extraction (87.8% accuracy in MedRU), performance plummets in patient-facing scenarios (61.3% in SmartServ). Moreover, the exceptional safety score (90.6% in MedSE) of the much smaller Baichuan-M2-32B highlights that targeted training is equally critical. Our specialized judge model, trained via SFT on a 19k expert-annotated medical dataset, achieves 92.1% accuracy, an F1-score of 94.37%, and a Cohen's Kappa of 81.3% for human-AI consistency, validating a reproducible and expert-aligned evaluation protocol. MLB thus provides a rigorous framework to guide the development of clinically viable LLMs.

cs.LG↗

arXiv2026

MedDialogRubrics: A Comprehensive Benchmark and Evaluation Framework for Multi-turn Medical Consultations in Large Language Models

Medical conversational AI (AI) plays a pivotal role in the development of safer and more effective medical dialogue systems. However, existing benchmarks and evaluation frameworks for assessing the information-gathering and diagnostic reasoning abilities of medical large language models (LLMs) have not been rigorously evaluated. To address these gaps, we present MedDialogRubrics, a novel benchmark comprising 5,200 synthetically constructed patient cases and over 60,000 fine-grained evaluation rubrics generated by LLMs and subsequently refined by clinical experts, specifically designed to assess the multi-turn diagnostic capabilities of LLM. Our framework employs a multi-agent system to synthesize realistic patient records and chief complaints from underlying disease knowledge without accessing real-world electronic health records, thereby mitigating privacy and data-governance concerns. We design a robust Patient Agent that is limited to a set of atomic medical facts and augmented with a dynamic guidance mechanism that continuously detects and corrects hallucinations throughout the dialogue, ensuring internal coherence and clinical plausibility of the simulated cases. Furthermore, we propose a structured LLM-based and expert-annotated rubric-generation pipeline that retrieves Evidence-Based Medicine (EBM) guidelines and utilizes the reject sampling to derive a prioritized set of rubric items ("must-ask" items) for each case. We perform a comprehensive evaluation of state-of-the-art models and demonstrate that, across multiple assessment dimensions, current models face substantial challenges. Our results indicate that improving medical dialogue will require advances in dialogue management architectures, not just incremental tuning of the base-model.

cs.CL↗

arXiv2025

Perplexity-Aware Data Scaling Law: Perplexity Landscapes Predict Performance for Continual Pre-training

Continual Pre-training (CPT) serves as a fundamental approach for adapting foundation models to domain-specific applications. Scaling laws for pre-training define a power-law relationship between dataset size and the test loss of an LLM. However, the marginal gains from simply increasing data for CPT diminish rapidly, yielding suboptimal data utilization and inefficient training. To address this challenge, we propose a novel perplexity-aware data scaling law to establish a predictive relationship between the perplexity landscape of domain-specific data and the test loss. Our approach leverages the perplexity derived from the pre-trained model on domain data as a proxy for estimating the knowledge gap, effectively quantifying the informational perplexity landscape of candidate training samples. By fitting this scaling law across diverse perplexity regimes, we enable adaptive selection of high-utility data subsets, prioritizing content that maximizes knowledge absorption while minimizing redundancy and noise. Extensive experiments demonstrate that our method consistently identifies near-optimal training subsets and achieves superior performance on both medical and general-domain benchmarks.

cs.LG↗

arXiv2025

GAPS: A Clinically Grounded, Automated Benchmark for Evaluating AI Clinicians

Current benchmarks for AI clinician systems, often based on multiple-choice exams or manual rubrics, fail to capture the depth, robustness, and safety required for real-world clinical practice. To address this, we introduce the GAPS framework, a multidimensional paradigm for evaluating Grounding (cognitive depth), Adequacy (answer completeness), Perturbation (robustness), and Safety. Critically, we developed a fully automated, guideline-anchored pipeline to construct a GAPS-aligned benchmark end-to-end, overcoming the scalability and subjectivity limitations of prior work. Our pipeline assembles an evidence neighborhood, creates dual graph and tree representations, and automatically generates questions across G-levels. Rubrics are synthesized by a DeepResearch agent that mimics GRADE-consistent, PICO-driven evidence review in a ReAct loop. Scoring is performed by an ensemble of large language model (LLM) judges. Validation confirmed our automated questions are high-quality and align with clinician judgment (90% agreement, Cohen's Kappa 0.77). Evaluating state-of-the-art models on the benchmark revealed key failure modes: performance degrades sharply with increased reasoning depth (G-axis), models struggle with answer completeness (A-axis), and they are highly vulnerable to adversarial perturbations (P-axis) as well as certain safety issues (S-axis). This automated, clinically-grounded approach provides a reproducible and scalable method for rigorously evaluating AI clinician systems and guiding their development toward safer, more reliable clinical practice. The benchmark dataset GAPS-NSCLC-preview and evaluation code are publicly available at https://huggingface.co/datasets/AQ-MedAI/GAPS-NSCLC-preview and https://github.com/AQ-MedAI/MedicalAiBenchEval.

cs.CL↗

arXiv2025

FunReason: Enhancing Large Language Models' Function Calling via Self-Refinement Multiscale Loss and Automated Data Refinement

The integration of large language models (LLMs) with function calling has emerged as a crucial capability for enhancing their practical utility in real-world applications. However, effectively combining reasoning processes with accurate function execution remains a significant challenge. Traditional training approaches often struggle to balance the detailed reasoning steps with the precision of function calls, leading to suboptimal performance. To address these limitations, we introduce FunReason, a novel framework that enhances LLMs' function calling capabilities through an automated data refinement strategy and a Self-Refinement Multiscale Loss (SRML) approach. FunReason leverages LLMs' natural reasoning abilities to generate high-quality training examples, focusing on query parseability, reasoning coherence, and function call precision. The SRML approach dynamically balances the contribution of reasoning processes and function call accuracy during training, addressing the inherent trade-off between these two critical aspects. FunReason achieves performance comparable to GPT-4o while effectively mitigating catastrophic forgetting during fine-tuning. FunReason provides a comprehensive solution for enhancing LLMs' function calling capabilities by introducing a balanced training methodology and a data refinement pipeline. For code and dataset, please refer to our repository at GitHub https://github.com/BingguangHao/FunReason

cs.LG↗