Xuchen Li

Xuchen Li, Shiyu Hu, Xiaokun Feng et al.

Visual Language Tracking (VLT) enhances tracking by mitigating the limitations of relying solely on the visual modality, utilizing high-level semantic information through language. This integration of the language enables more advanced human-machine interaction. The essence of interaction is cognitive alignment, which typically requires multiple information exchanges, especially in the sequential decision-making process of VLT. However, current VLT benchmarks do not account for multi-round interactions during tracking. They provide only an initial text and bounding box (bbox) in the first frame, with no further interaction as tracking progresses, deviating from the original motivation of the VLT task. To address these limitations, we propose a novel and robust benchmark, VLT-MI (Visual Language Tracking with Multi-modal Interaction), which introduces multi-round interaction into the VLT task for the first time. (1) We generate diverse, multi-granularity texts for multi-round, multi-modal interaction based on existing mainstream VLT benchmarks using DTLLM-VLT, leveraging the world knowledge of LLMs. (2) We propose a new VLT interaction paradigm that achieves multi-round interaction through text updates and object recovery. When multiple tracking failures occur, we provide the tracker with more aligned texts and corrected bboxes through interaction, thereby expanding the scope of VLT downstream tasks. (3) We conduct comparative experiments on both traditional VLT benchmarks and VLT-MI, evaluating and analyzing the accuracy and robustness of trackers under the interactive paradigm. This work offers new insights and paradigms for the VLT task, enabling a fine-grained evaluation of multi-modal trackers. We believe this approach can be extended to additional datasets in the future, supporting broader evaluations and comparisons of video-language model capabilities.

Xuchen Li, Xuzhao Li, Shiyu Hu et al.

Recent advances in large language models (LLMs) have improved reasoning in text and image domains, yet achieving robust video reasoning remains a significant challenge. Existing video benchmarks mainly assess shallow understanding and reasoning and allow models to exploit global context, failing to rigorously evaluate true causal and stepwise reasoning. We present CausalStep, a benchmark designed for explicit stepwise causal reasoning in videos. CausalStep segments videos into causally linked units and enforces a strict stepwise question-answer (QA) protocol, requiring sequential answers and preventing shortcut solutions. Each question includes carefully constructed distractors based on error type taxonomy to ensure diagnostic value. The benchmark features 100 videos across six categories and 1,852 multiple-choice QA pairs. We introduce seven diagnostic metrics for comprehensive evaluation, enabling precise diagnosis of causal reasoning capabilities. Experiments with leading proprietary and open-source models, as well as human baselines, reveal a significant gap between current models and human-level stepwise reasoning. CausalStep provides a rigorous benchmark to drive progress in robust and interpretable video reasoning.

Xuchen Li, Xuzhao Li, Shiyu Hu et al.

Long-form video reasoning remains a major challenge for Video Large Language Models (Video LLMs), as static uniform frame sampling leads to information dilution and obscures critical evidence. Furthermore, existing pixel-space video reasoning agents, which are designed to actively interact with the video to acquire new visual information, remain suboptimal due to their lack of rigorous reward mechanisms to enforce evidence purity and their inability to perform temporal information supplementation beyond pre-sampled frames. To address this critical gap, we propose a novel evidence-prioritized adaptive framework built upon our core philosophy: "Select Less, Reason More." Our core contribution is the evidence-aware reinforcement learning (EARL) framework, which transforms the model into an active interrogator of evidence. EARL is precisely engineered to dynamically select the most relevant frames and, crucially, to perform localized re-sampling around the selected key frames to access fine-grained temporal detail. Extensive experiments on five demanding video reasoning benchmarks demonstrate that our EARL-trained model achieves new state-of-the-art among open-source Video LLMs, simultaneously learning an effective and high-purity visual evidence selection policy. Impressively, our 7B model achieves 59.8% on LongVideoBench, 69.0% on MVBench and 64.9% on VideoMME. These results highlight the importance of prioritizing evidence purity and the effectiveness of our framework.

Xuchen Li, Xiaokun Feng, Shiyu Hu et al.

Visual Language Tracking (VLT) enhances single object tracking (SOT) by integrating natural language descriptions from a video, for the precise tracking of a specified object. By leveraging high-level semantic information, VLT guides object tracking, alleviating the constraints associated with relying on a visual modality. Nevertheless, most VLT benchmarks are annotated in a single granularity and lack a coherent semantic framework to provide scientific guidance. Moreover, coordinating human annotators for high-quality annotations is laborious and time-consuming. To address these challenges, we introduce DTLLM-VLT, which automatically generates extensive and multi-granularity text to enhance environmental diversity. (1) DTLLM-VLT generates scientific and multi-granularity text descriptions using a cohesive prompt framework. Its succinct and highly adaptable design allows seamless integration into various visual tracking benchmarks. (2) We select three prominent benchmarks to deploy our approach: short-term tracking, long-term tracking, and global instance tracking. We offer four granularity combinations for these benchmarks, considering the extent and density of semantic information, thereby showcasing the practicality and versatility of DTLLM-VLT. (3) We conduct comparative experiments on VLT benchmarks with different text granularities, evaluating and analyzing the impact of diverse text on tracking performance. Conclusionally, this work leverages LLM to provide multi-granularity semantic information for VLT task from efficient and diverse perspectives, enabling fine-grained evaluation of multi-modal trackers. In the future, we believe this work can be extended to more datasets to support vision datasets understanding.

Xuchen Li, Shiyu Hu, Xiaokun Feng et al.

Vision-language tracking (VLT) extends traditional single object tracking by incorporating textual information, providing semantic guidance to enhance tracking performance under challenging conditions like fast motion and deformations. However, current VLT trackers often underperform compared to single-modality methods on multiple benchmarks, with semantic information sometimes becoming a "distraction." To address this, we propose VLTVerse, the first fine-grained evaluation framework for VLT trackers that comprehensively considers multiple challenge factors and diverse semantic information, hoping to reveal the role of language in VLT. Our contributions include: (1) VLTVerse introduces 10 sequence-level challenge labels and 6 types of multi-granularity semantic information, creating a flexible and multi-dimensional evaluation space for VLT; (2) leveraging 60 subspaces formed by combinations of challenge factors and semantic types, we conduct systematic fine-grained evaluations of three mainstream SOTA VLT trackers, uncovering their performance bottlenecks across complex scenarios and offering a novel perspective on VLT evaluation; (3) through decoupled analysis of experimental results, we examine the impact of various semantic types on specific challenge factors in relation to different algorithms, providing essential guidance for enhancing VLT across data, evaluation, and algorithmic dimensions. The VLTVerse, toolkit, and results will be available at \url{http://metaverse.aitestunion.com}.

Yiping Ma, Shiyu Hu, Xuchen Li et al.

Recent advances in large language models (LLMs) have enabled intelligent tutoring systems, yet the development of LLM-based Virtual Student Agents (LVSAs) remains underexplored. Such agents are essential for teacher-facing applications, where simulating diverse learner traits can support adaptive instruction and pedagogical skill development. However, current methods lack principled personality modeling, scalable evaluation of behavioral consistency, and empirical validation in interactive teaching settings. We propose the SOEI framework, a structured pipeline comprising Scene, Object, Evaluation, and Interaction, for constructing and evaluating personality-aligned LVSAs in classroom scenarios. Leveraging Chinese language instruction as a cognitively and emotionally rich testbed, we generate five LVSAs based on Big Five traits through LoRA fine-tuning and expert-informed prompt design. Their behavioral realism and personality coherence are assessed using a hybrid human & GPT-4 evaluation and a multi-dimensional annotation protocol. Through controlled experiments with real pre-service teachers, we demonstrate that LVSAs can elicit adaptive teaching strategies and maintain trait-consistent behavior across multi-turn dialogues. Our results provide: (1) an educationally and psychologically grounded generation pipeline for LLM-based student agents; (2) a hybrid, scalable evaluation framework for behavioral realism; and (3) empirical insights into the pedagogical utility of LVSAs in shaping instructional adaptation. By embedding LVSAs into both generative modeling and human-in-the-loop teaching, SOEI bridges AI for Education (AI4Edu) and Education for AI (Edu4AI), positioning classroom interaction as a rigorous testbed for controllability, personality alignment, and human-likeness in large language models.

Xuzhao Li, Xuchen Li, Shiyu Hu

Nighttime UAV tracking presents significant challenges due to extreme illumination variations and viewpoint changes, which severely degrade tracking performance. Existing approaches either rely on light enhancers with high computational costs or introduce redundant domain adaptation mechanisms, failing to fully utilize the dynamic features in varying perspectives. To address these issues, we propose \textbf{DARTer} (\textbf{D}ynamic \textbf{A}daptive \textbf{R}epresentation \textbf{T}racker), an end-to-end tracking framework designed for nighttime UAV scenarios. DARTer leverages a Dynamic Feature Blender (DFB) to effectively fuse multi-perspective nighttime features from static and dynamic templates, enhancing representation robustness. Meanwhile, a Dynamic Feature Activator (DFA) adaptively activates Vision Transformer layers based on extracted features, significantly improving efficiency by reducing redundant computations. Our model eliminates the need for complex multi-task loss functions, enabling a streamlined training process. Extensive experiments on multiple nighttime UAV tracking benchmarks demonstrate the superiority of DARTer over state-of-the-art trackers. These results confirm that DARTer effectively balances tracking accuracy and efficiency, making it a promising solution for real-world nighttime UAV tracking applications.

Shiyu Hu, Xuchen Li, Xuzhao Li et al.

Despite rapid progress in large vision-language models (LVLMs), existing video caption benchmarks remain limited in evaluating their alignment with human understanding. Most rely on a single annotation per video and lexical similarity-based metrics, failing to capture the variability in human perception and the cognitive importance of events. These limitations hinder accurate diagnosis of model capabilities in producing coherent, complete, and human-aligned descriptions. To address this, we introduce FIOVA (Five-In-One Video Annotations), a human-centric benchmark tailored for evaluation. It comprises 3,002 real-world videos (about 33.6s each), each annotated independently by five annotators. This design enables modeling of semantic diversity and inter-subjective agreement, offering a richer foundation for measuring human-machine alignment. We further propose FIOVA-DQ, an event-level evaluation metric that incorporates cognitive weights derived from annotator consensus, providing fine-grained assessment of event relevance and semantic coverage. Leveraging FIOVA, we conduct a comprehensive evaluation of nine representative LVLMs and introduce a complexity-aware analysis framework based on inter-annotator variation (CV). This reveals consistency gaps across difficulty levels and identifies structural issues such as event under-description and template convergence. Our results highlight FIOVA's diagnostic value for understanding LVLM behavior under varying complexity, setting a new standard for cognitively aligned evaluation in long-video captioning. The benchmark, annotations, metric, and model outputs are publicly released to support future evaluation-driven research in video understanding. More detailed information can be found at https://huuuuusy.github.io/fiova/.

Xuchen Li, Xuzhao Li, Jiahui Gao et al.

Vision-Language Models (VLMs) excel at many multimodal tasks, yet they frequently struggle with tasks requiring precise understanding and handling of fine-grained visual elements. This is mainly due to information loss during image encoding or insufficient attention to critical regions. Recent work has shown promise by incorporating pixel-level visual information into the reasoning process, enabling VLMs to access high-resolution visual details during their thought process. However, this pixel-level information is often overused, leading to inefficiency and distraction from irrelevant visual details. To address these challenges, we propose the first framework for adaptive pixel reasoning that dynamically determines necessary pixel-level operations based on the input query. Specifically, we first apply operation-aware supervised fine-tuning to establish baseline competence in textual reasoning and visual operations, then design a novel rollout-guided reinforcement learning framework relying on feedback of the model's own responses, which enables the VLM to determine when pixel operations should be invoked based on query difficulty. Experiments on extensive multimodal reasoning benchmarks show that our model achieves superior performance while significantly reducing unnecessary visual operations. Impressively, our model achieves 73.4\% accuracy on HR-Bench 4K while maintaining a tool usage ratio of only 20.1\%, improving accuracy and simultaneously reducing tool usage by 66.5\% compared to the previous methods.