Dongsheng Li

34

Papers

126

Total Citations

Papers (34)

Disentangled Prompt Representation for Domain Generalization

Mitigating Hallucinations in Large Vision-Language Models via DPO: On-Policy Data Hold the Key

Auto-Prox: Training-Free Vision Transformer Architecture Search via Automatic Proxy Discovery

CNN Kernels Can Be the Best Shapelets

DreamDistribution: Learning Prompt Distribution for Diverse In-distribution Generation

Unified Medical Image Pre-training in Language-Guided Common Semantic Space

Learning to Instruct for Visual Instruction Tuning

NeurIPS 2025arXiv

Oracle-MoE: Locality-preserving Routing in the Oracle Space for Memory-constrained Large Language Model Inference

Rotation-Invariant Transformer for Point Cloud Matching

Correlation Congruence for Knowledge Distillation

Energy-Based Open-World Uncertainty Modeling for Confidence Calibration

Mixture-Rank Matrix Approximation for Collaborative Filtering

Towards Precise End-to-End Weakly Supervised Object Detection Network

Adversarial Domain Prompt Tuning and Generation for Single Domain Generalization

VisRL: Intention-Driven Visual Perception via Reinforced Reasoning

Zero-resource Hallucination Detection for Text Generation via Graph-based Contextual Knowledge Triples Modeling

Learning Hierarchical Prompt with Structured Linguistic Knowledge for Vision-Language Models

Efficient and Effective Time-Series Forecasting with Spiking Neural Networks

Learning Invariant Representations and Risks for Semi-Supervised Domain Adaptation

RendNet: Unified 2D/3D Recognizer With Latent Space Rendering

General Proximal Incremental Aggregated Gradient Algorithms: Better and Novel Results under General Scheme

Reinforcement Learning Enhanced Explainer for Graph Neural Networks

Recognizing Vector Graphics without Rasterization

Reinforcement Learning with Automated Auxiliary Loss Search

Transcormer: Transformer for Sentence Scoring with Sliding Language Modeling

Finite-Time Analysis of Adaptive Temporal Difference Learning with Deep Neural Networks

Parameter-free Dynamic Graph Embedding for Link Prediction

VRL3: A Data-Driven Framework for Visual Deep Reinforcement Learning

Bootstrapped Transformer for Offline Reinforcement Learning

Train Faster, Perform Better: Modular Adaptive Training in Over-Parameterized Models

HuggingGPT: Solving AI Tasks with ChatGPT and its Friends in Hugging Face

ContiFormer: Continuous-Time Transformer for Irregular Time Series Modeling

Learning Topology-Agnostic EEG Representations with Geometry-Aware Modeling

Low-Rank Matrix Approximation with Stability