16 papers
Control-Diverse Reinforcement Fine-Tuning: Decoupling the Shared Control Bottleneck of RL Post-Training
Binwen Tan, Jingchao Wang, Dengzhe Hou +6
Reinforcement learning post-training unlocks complex reasoning in LLMs. Yet benchmark scores reveal only whether a model improved, not what changed inside it, nor how it splits fin…
CogEEGAgent: Toward Autonomous Cognitive EEG Analysis with Grounded Execution and Selection-Aware Verification
Dengzhe Hou, Lingyu Jiang, Fangzhou Lin +1
Electroencephalography (EEG) analysis in cognitive studies requires specialized expertise and involves many defensible choices over contrasts, channels, time windows, and statistic…
CV-Arena: An Open Benchmark for Instructional Computer Vision Problem Solving with Human-AI Collaborative Preferences
Fangzhou Lin, Peiran Li, Lingyu Xu +12
Instruction-guided image editing is becoming a general interface for visual work, yet existing benchmarks still focus largely on narrow appearance edits and do not fully capture th…
PathCal: State-Aware Reflection-Marker Calibration for Efficient Reasoning
Lingyu Jiang, Zirui Li, Shuo Xing +7
Large Reasoning Language Models (LRMs) tackle complex tasks through test-time scaling, generating long-form Chain-of-Thought (CoT) trajectories during inference. These trajectories…
CAPS: Cascaded Adaptive Pairwise Selection for Efficient Parallel Reasoning
Fangzhou Lin, Shuo Xing, Peiran Li +6
Parallel reasoning, where a generator samples many candidate solutions and an aggregator selects the best, is one of the most effective forms of test-time scaling in large language…
Physics-Aware Video Instance Removal Benchmark
Zirui Li, Xinghao Chen, Lingyu Jiang +5
Video Instance Removal (VIR) requires removing target objects while maintaining background integrity and physical consistency, such as specular reflections and illumination interac…
AdaptFuse: Training-Free Sequential Preference Inference via Externalized Bayesian Computation
Fangzhou Lin, Zilin Dai, Peiran Li +8
Large language models (LLMs) struggle to accumulate evidence across multiple rounds of user interactions, failing to update their beliefs in a manner consistent with Bayesian infer…
Let the Abyss Stare Back Adaptive Falsification for Autonomous Scientific Discovery
Peiran Li, Fangzhou Lin, Shuo Xing +5
Autonomous scientific discovery is entering a more dangerous regime: once the evaluator is frozen, a sufficiently strong search process can learn to win the exam without learning t…
The Pulse of Motion: Measuring Physical Frame Rate from Visual Dynamics
Xiangbo Gao, Mingyang Wu, Siyuan Yang +4
While recent generative video models have achieved remarkable visual realism and are being explored as world models, true physical simulation requires mastering both space and time…
Industrial3D: A Water-Treatment TLS Point Cloud Dataset and Cross-Paradigm Benchmark for MEP Scene Understanding
Chao Yin, Hongzhe Yue, Qing Han +8
Automated semantic understanding of dense terrestrial laser scanning (TLS) point clouds is a prerequisite for Scan-to-BIM, digital twin maintenance, and as-built verifcation. Yet f…
Position: Human-Centric AI Requires a Minimum Viable Level of Human Understanding
Fangzhou Lin, Qianwen Ge, Lingyu Xu +7
AI systems increasingly produce fluent, correct, end-to-end outcomes. Over time, this erodes users' ability to explain, verify, or intervene. We define this divergence as the Capab…
Traversal-as-Policy: Log-Distilled Gated Behavior Trees as Externalized, Verifiable Policies for Safe, Robust, and Efficient Agents
Peiran Li, Jiashuo Sun, Fangzhou Lin +5
Autonomous LLM agents fail because long-horizon policy remains implicit in model weights and transcripts, while safety is retrofitted post hoc. We propose Traversal-as-Policy: dist…
BibAgent: An Agentic Framework for Traceable Miscitation Detection in Scientific Literature
Peiran Li, Fangzhou Lin, Shuo Xing +6
Citations are the bedrock of scientific authority, yet their integrity is compromised by widespread miscitations: ranging from nuanced distortions to fabricated references. Systema…
NexusFlow: Unifying Disparate Tasks under Partial Supervision via Invertible Flow Networks
Fangzhou Lin, Yuping Wang, Yuliang Guo +7
Partially Supervised Multi-Task Learning (PS-MTL) aims to leverage knowledge across tasks when annotations are incomplete. Existing approaches, however, have largely focused on the…
TimePre: Bridging Accuracy, Efficiency, and Stability in Probabilistic Time-Series Forecasting
Lingyu Jiang, Lingyu Xu, Peiran Li +14
We propose TimePre, a simple framework that unifies the efficiency of Multilayer Perceptron (MLP)-based models with the distributional flexibility of Multiple Choice Learning (MCL)…
SafeCoop: Unravelling Full Stack Safety in Agentic Collaborative Driving
Xiangbo Gao, Tzu-Hsiang Lin, Ruojing Song +6
Collaborative driving systems leverage vehicle-to-everything (V2X) communication across multiple agents to enhance driving safety and efficiency. Traditional V2X systems take raw s…