Yu Liu

AdPO: Enhancing the Adversarial Robustness of Large Vision-Language Models with Preference Optimization

FACM: Flow-Anchored Consistency Models

Towards Sequence Modeling Alignment between Tokenizer and Autoregressive Model

DenseGRPO: From Sparse to Dense Reward for Flow Matching Model Alignment

Routing Matters in MoE: Scaling Diffusion Transformers with Explicit Routing Guidance

SRDiffusion: Accelerate Diffusion Inference via Sketching-Rendering Cooperation

Value-Anchored Group Policy Optimization for Flow Models

Instability in Diffusion ODEs: An Explanation for Inaccurate Image Reconstruction

Wan-Move: Motion-controllable Video Generation via Latent Trajectory Guidance

The Matrix: Infinite-Horizon World Generation with Real-Time Moving Control

ViewPoint: Panoramic Video Generation with Pretrained Diffusion Models

ACE: All-round Creator and Editor Following Instructions via Diffusion Transformer

MiCo: Multi-image Contrast for Reinforcement Visual Reasoning

Asking Specifically Instead of Ambiguously to Your GPT Improves Image Caption

Improved Video VAE for Latent Video Diffusion Model

Group Diffusion Transformers are Unsupervised Multitask Learners

AdPO: Enhancing the Adversarial Robustness of Large Vision-Language Models with Preference Optimization