Dong Li

27

Papers

79

Total Citations

1

Affiliations

Affiliations

Shanghai AI Lab

Papers (27)

Interactive Continual Learning: Fast and Slow Thinking

CO2: Efficient Distributed Training with Full Communication-Computation Overlap

Noise-assisted Prompt Learning for Image Forgery Detection and Localization

Týr-the-Pruner: Structural Pruning LLMs via Global Sparsity Distribution Optimization

NeurIPS 2025arXiv

ReNeg: Learning Negative Embedding with Reward Guidance

UHD-processer: Unified UHD Image Restoration with Progressive Frequency Learning and Degradation-aware Prompts

Multiagent Gumbel MuZero: Efficient Planning in Combinatorial Action Spaces

CoE: Chain-of-Explanation via Automatic Visual Concept Circuit Description and Polysemanticity Quantification

SolverLLM: Leveraging Test-Time Scaling for Optimization Problem via LLM-Guided Search

Fast and Slow Gradient Approximation for Binary Neural Network Optimization

PG-LBO: Enhancing High-Dimensional Bayesian Optimization with Pseudo-Label and Gaussian Process Guidance

DR-Label: Label Deconstruction and Reconstruction of GNN Models for Catalysis Systems

780 Learning Discriminative Noise Guidance for Image Forgery Detection and Localization

CatmullRom Splines-Based Regression for Image Forgery Localization

HomoFormer: Homogenized Transformer for Image Shadow Removal

Improving Unsupervised Hierarchical Representation with Reinforcement Learning

HarmonyDream: Task Harmonization Inside World Models

Various Lengths, Constant Speed: Efficient Language Modeling with Lightning Attention

Enhancing Vision Transformer: Amplifying Non-Linearity in Feedforward Network Module

Pursuing Temporal-Consistent Video Virtual Try-On via Dynamic Pose Interaction

Less is More: Efficient Model Merging with Binary Task Switch

Decouple to Reconstruct: High Quality UHD Restoration via Active Feature Disentanglement and Reversible Fusion

FreeDNA: Endowing Domain Adaptation of Diffusion-Based Dense Prediction with Training-Free Domain Noise Alignment

Enhanced Pansharpening via Quaternion Spatial-Spectral Interactions

EGSRAL:An Enhanced 3D Gaussian Splatting Based Renderer with Automated Labeling for Large-Scale Driving Scene

DreamUHD: Frequency Enhanced Variational Autoencoder for Ultra-High-Definition Image Restoration

A Lottery Ticket Hypothesis Approach with Sparse Fine-tuning and MAE for Image Forgery Detection and Localization