11 papers
SpikeOcc: Towards Resilient Multimodal 3D Occupancy Perception via Spiking Experts
Yiding Sun, Zijie Xu, Hongqiang Lin +7
Multimodal 3D occupancy prediction underpins safe autonomous driving through comprehensive scene understanding. However, existing methods have largely pursued higher accuracy throu…
Beyond Endpoint Performance: Process-Level Evaluation of Self-Evolving Agents
Hongqiang Lin, Chao Liu, Xiaofan Bai +4
Self-evolving agents convert interaction feedback into persistent artifacts, such as memories or skills, which in turn guide subsequent decisions. As these artifacts are iterativel…
SkillZip Pro: Execution-Aware Dynamic Compression of Progressively Loaded Skills for Self-Evolving Agents
Xiaofan Bai, Chao Liu, Hongqiang Lin +5
Production agent skills are directory bundles, not isolated prompts. The root is loaded at activation; references, schemas, scripts, assets, and nested subskills are loaded only wh…
Rethinking Self-Evolution: A Constrained Exploration-Exploitation Process for Mitigating Skill Overfitting
Hongqiang Lin, Chao Liu, Xiaofan Bai +4
Enabling large language model (LLM) agents to accumulate and reuse experience from past interactions remains a central challenge in real-world applications. A promising solution is…
SkillZip: Evaluation-Free Skill Compression for Self-Evolving Agents by Discovering Reusable Structure
Xiaofan Bai, Hongqiang Lin, Chao Liu +4
Self-evolving agents accumulate reusable skills by appending successful procedures and failure fixes. Over time, the same requirement is often restated in several branches, example…
Regularized Offline Policy Optimization with Posterior Hybrid Bayesian Belief
Hongqiang Lin, Pengfei Wang, Nenggan Zheng
Offline reinforcement learning (RL) aims to optimize policies from pre-collected datasets. A bottleneck of this paradigm is managing epistemic uncertainty, which arises from limite…
Offline Policy Optimization with Posterior Sampling
Hongqiang Lin, Dongxu Zhang, Yiding Sun +3
A fundamental challenge in model-based offline reinforcement learning (RL) lies in the trade-off between generalization and robustness against exploitation errors in out-of-distrib…
CFMS: A Coarse-to-Fine Multimodal Synthesis Framework for Enhanced Tabular Reasoning
Qixian Huang, Hongqiang Lin, Tong Fu +5
Reasoning over tabular data is a crucial capability for tasks like question answering and fact verification, as it requires models to comprehend both free-form questions and semi-s…
Not All Queries Need Deep Thought: CoFiCot for Adaptive Coarse-to-fine Stateful Refinement
Dongxu Zhang, Hongqiang Lin, Yiding Sun +4
Scaling test-time computation enhances LLM reasoning ability but faces a uniform computation paradox. Allocating identical resources leads to over-correction on simple tasks and in…
Robust Regularized Policy Iteration under Transition Uncertainty
Hongqiang Lin, Zhenghui Fu, Weihao Tang +4
Offline reinforcement learning (RL) enables data-efficient and safe policy learning without online exploration, but its performance often degrades under distribution shift. The lea…
PointCoT: A Multi-modal Benchmark for Explicit 3D Geometric Reasoning
Dongxu Zhang, Yiding Sun, Pengcheng Li +12
While Multimodal Large Language Models (MLLMs) demonstrate proficiency in 2D scenes, extending their perceptual intelligence to 3D point cloud understanding remains a significant c…