Hongtao Xie

Unsupervised Reinforcement Learning with Verifiable Rewards via First Repeat Criterion

Visual Feedback for Self-Improving Text Layout with MLLM via Reinforcement Learning

From Evaluation to Defense: Advancing Safety in Video Large Language Models

SDErasure: Concept-Specific Trajectory Shifting for Concept Erasure via Adaptive Diffusion Classifier

Test-Time Scaling with Reflective Generative Model

UpSafe℃: Upcycling for Controllable Safety in Large Language Models

One-for-All: Towrads Human-Centric Multi-Subject Customization from Single-Subject Examples

INPO: Image-based Negative Preference Optimization for Concept Erasure in Text-to-Image Diffusion Models

A Hybrid Paradigm for Vision Autoencoders: Unifying CNNs and Transformers for Learning Efficiency and Scalability

Latents-Inv:Robust Semantic Watermark with Key-Assisted Recovery for diffusion models

GRIP: A Graph-Based Reasoning Instruction Producer

A Graph-Based Synthetic Data Pipeline for Scaling High-Quality Data