Qinghua Hu

Mixture of Groups: Grouped Gating and Cross Mixing for Parameter-Efficient LLM Fine-Tuning

Dig2DIG: Dig into Diffusion Information Gains for Image Fusion

Adapting LLMs to Time Series Forecasting via Temporal Heterogeneity Modeling and Semantic Alignment

PE-DyRA: Dynamic Rank Adaptation for Parameter-Efficient Fine-Tuning via Importance-Aware Pruning and Expansion

Semantic Energy: Detecting LLM Hallucination Beyond Entropy

Local Uncertainty Smoothing For Few-shot OOD Detection

Multimodal Negative Learning

Graphs Help Graphs: Multi-Agent Graph Socialized Learning

Task-Gated Multi-Expert Collaboration Network for Degraded Multi-Modal Image Fusion

RoomEditor: High-Fidelity Furniture Synthesis with Parameter-Sharing U-Net

Asymmetric Factorized Bilinear Operation for Vision Transformer

Efficient Masked AutoEncoder for Video Object Counting and A Large-Scale Benchmark

Patch-wise Structural Loss for Time Series Forecasting

DOTA: Distributional Test-Time Adaptation of Vision-Language Models

Learning Pattern-Specific Experts for Time Series Forecasting Under Patch-level Distribution Shift

Online Clustering of Dueling Bandits

Socialized Coevolution: Advancing a Better World through Cross-Task Collaboration

Learning Pattern-Specific Experts for Time Series Forecasting Under Patch-level Distribution Shift

Towards Multimodal Open Set Recognition

Reducing class-wise confusion for incremental learning with disentangled manifolds