Harry Yang

8

Papers

57

Total Citations

Papers (8)

Intervening Anchor Token: Decoding Strategy in Alleviating Hallucinations for MLLMs

Unveiling the Ignorance of MLLMs: Seeing Clearly, Answering Incorrectly

DreamDance: Animating Human Images by Enriching 3D Geometry Cues from 2D Poses

Model Reveals What to Cache: Profiling-Based Feature Reuse for Video Diffusion Models

Robustness and Generalization via Generative Adversarial Training

MUGEN: A Playground for Video-Audio-Text Multimodal Understanding and GENeration

Long Video Generation with Time-Agnostic VQGAN and Time-Sensitive Transformer

Using Mixup as a Regularizer Can Surprisingly Improve Accuracy & Out-of-Distribution Robustness