21 papers
GeoNest: Learning to Select Failure-Aware Neighborhoods for the Irregular Knapsack Problem in a Circular Container
Zhongman Du, Huiming Zhang, Linlin Yang +2
The two-dimensional irregular knapsack problem in a fixed circular container is an important combinatorial optimization problem for maximizing material utilization in manufacturing…
GleanVID: Complementary Token Selection for Efficient Video Large Language Models
Shuo Yang, Changbai Li, Rui Tang +3
Video Large Language Models (VideoLLMs) have achieved strong video understanding capabilities but incur substantial inference overhead due to the large number of visual tokens. Exi…
GraphHCA: Closed-Form Hindsight Credit Assignment for Long-Horizon LLM Agents
Haodong Zhu, Yangyang Ren, Changbai Li +4
Group-based reinforcement learning (RL) has advanced large language models (LLMs) and is increasingly extending to agentic tasks, where sparse terminal rewards make step-level cred…
Cross-Rollout Bellman Closure for Long-Horizon Agentic Reinforcement Learning
Yangyang Ren, Haodong Zhu, Linlin Yang +3
Group-based reinforcement learning such as GRPO trains LLM agents by comparing rollouts sampled for each task, without a learned critic. In long-horizon settings, these rollouts re…
MHE-Former: Multi-Hypothesis Transformers via Entropy Maximization for 3D Mesh Recovery
Boshu Jia, Rongyu Chen, Linlin Yang +9
Monocular 3D hand and body mesh recovery often suffers from severe occlusion and ambiguity. Traditional deterministic methods typically regress a single optimal solution, leading t…
Contextual Information Policy Optimization for Search Agents
Xingyu Guo, Wei Chen, Linlin Yang +1
Search agents extend large language models beyond static parametric memory by enabling them to acquire and use external evidence during multi-step reasoning. For knowledge-intensiv…
Kalman Meets Curriculum: Efficient Dynamic Prompt Selection for Adaptive RL Finetuning
Haodong Zhu, Yangyang Ren, Yanjing Li +4
Reinforcement learning (RL) finetuning significantly enhances the reasoning capabilities of large language models (LLMs), yet its effectiveness critically depends on selecting prom…
FAIR-Calib: Frontier-Aware Instability-Reweighted Calibration for Post-Training Quantization of Diffusion Large Language Models
Haoyu Huang, Linlin Yang, Sheng Xu +5
Diffusion Large Language Models (dLLMs) refine tokens iteratively but commit them irreversibly, leading to a "stability lag" where early decisions remain fragile even after being w…
LDFE: Laplacian Decoupled Feature Enhancement Block for Dual-Stream CNN-based RGB-IR Object Detection
Wenhao Dong, Xiaoyan Luo, Linlin Yang +4
The complementary information between RGB and IR images can significantly enhance object detection performance under extreme conditions. Existing methods prefer dual-stream CNN bac…
InfraNet: Quality-Aware RGB Guidance for Efficient Infrared Object Detection
Zichao Feng, Haodong Zhu, Jingying Yang +8
Robust object detection under adverse visual conditions remains a long-standing challenge for multi-modal perception systems. Existing fusion-based methods typically require both R…
Teaching Vision-Language-Action Models What to See and Where to Look
Yuguang Yang, Canyu Chen, Zhewen Tan +10
Vision-Language-Action (VLA) models have emerged as a promising paradigm for end-to-end autonomous driving. However, existing VLAs' training relies heavily on text-centric visual q…
CL-CLIP: CLIP-Based Continual Learning Framework with Cost-Volume Category Decoupling for Object Detection
Zihan Liu, Yuguang Yang, Shengjie Su +5
Continual Object Detection (COD) requires a detector to acquire new categories over time while preserving previously learned ones. This goal is closely related to open-vocabulary d…
Devil is in Narrow Policy: Unleashing Exploration in Driving VLA Models
Canyu Chen, Yuguang Yang, Zhewen Tan +10
We identify a fundamental Narrow Policy limitation undermining the performance of autonomous VLA models, where driving Imitation Learning (IL) tends to collapse exploration and lim…
Unbiased Dynamic Pruning for Efficient Group-Based Policy Optimization
Haodong Zhu, Yangyang Ren, Yanjing Li +6
Group Relative Policy Optimization (GRPO) effectively scales LLM reasoning but incurs prohibitive computational costs due to its extensive group-based sampling requirement. While r…
UrbanGS: A Scalable and Efficient Architecture for Geometrically Accurate Large-Scene Reconstruction
Changbai Li, Haodong Zhu, Hanlin Chen +6
While 3D Gaussian Splatting (3DGS) enables high-quality, real-time rendering for bounded scenes, its extension to large-scale urban environments gives rise to critical challenges i…
Efficiently Seeking Flat Minima for Better Generalization in Fine-Tuning Large Language Models and Beyond
Jiaxin Deng, Qingcheng Zhu, Junbiao Pang +3
Little research explores the correlation between the expressive ability and generalization ability of the low-rank adaptation (LoRA). Sharpness-Aware Minimization (SAM) improves mo…
WaveMamba: Wavelet-Driven Mamba Fusion for RGB-Infrared Object Detection
Haodong Zhu, Wenhao Dong, Linlin Yang +10
Leveraging the complementary characteristics of visible (RGB) and infrared (IR) imagery offers significant potential for improving object detection. In this paper, we propose WaveM…
Prompt as Knowledge Bank: Boost Vision-language model via Structural Representation for zero-shot medical detection
Yuguang Yang, Tongfei Chen, Haoyu Huang +5
Zero-shot medical detection can further improve detection performance without relying on annotated medical images even upon the fine-tuned model, showing great clinical value. Rece…
Normalizing Batch Normalization for Long-Tailed Recognition
Yuxiang Bao, Guoliang Kang, Linlin Yang +3
In real-world scenarios, the number of training samples across classes usually subjects to a long-tailed distribution. The conventionally trained network may achieve unexpected inf…
DecomCAM: Advancing Beyond Saliency Maps through Decomposition and Integration
Yuguang Yang, Runtang Guo, Sheng Wu +6
Interpreting complex deep networks, notably pre-trained vision-language models (VLMs), is a formidable challenge. Current Class Activation Map (CAM) methods highlight regions revea…
Bias-Compensated Integral Regression for Human Pose Estimation
Kerui Gu, Linlin Yang, Michael Bi Mi +1
In human and hand pose estimation, heatmaps are a crucial intermediate representation for a body or hand keypoint. Two popular methods to decode the heatmap into a final joint coor…