8 papers
Correction Training for Diffusion Models
Donghoon Ahn, Runqian Wang, Alexei A. Efros +2
Iterative sampling in diffusion models can lead to error accumulation due to exposure bias. The denoiser is trained on real images interpolated with noise, but at inference its inp…
Equilibrium Forcing: Adaptive Video Generation Without Noise Conditioning
Hansen Jin Lillemark, Alex Rojas, Zachary Novack +5
Standard autoregressive video generation algorithms based on Diffusion and Flow Matching rely on rigid training objectives and static sampling schedules, limiting inference procedu…
Unlearning Offline Stochastic Multi-Armed Bandits
Zichun Ye, Runqi Wang, Xuchuang Wang +3
Machine unlearning aims to unlearn data points from a learned model, offering a principled way to process data-deletion requests and mitigate privacy risks without full retraining.…
ARC Is a Vision Problem!
Keya Hu, Ali Cy, Linlu Qiu +5
The Abstraction and Reasoning Corpus (ARC) is designed to promote research on abstract reasoning, a fundamental aspect of human intelligence. Common approaches to ARC treat it as a…
Equilibrium Matching: Generative Modeling with Implicit Energy-Based Models
Runqian Wang, Yilun Du
We introduce Equilibrium Matching (EqM), a generative modeling framework built from an equilibrium dynamics perspective. EqM discards the non-equilibrium, time-conditional dynamics…
Diffuse and Disperse: Image Generation with Representation Regularization
Runqian Wang, Kaiming He
The development of diffusion-based generative models over the past decade has largely proceeded independently of progress in representation learning. These diffusion models typical…
VLMInferSlow: Evaluating the Efficiency Robustness of Large Vision-Language Models as a Service
Xiasi Wang, Tianliang Yao, Simin Chen +5
Vision-Language Models (VLMs) have demonstrated great potential in real-world applications. While existing research primarily focuses on improving their accuracy, the efficiency re…
: towards data-free Transferable Parameter Efficient Finetuning
Runqian Wang, Soumya Ghosh, David Cox +4
Low-rank adapters (LoRA) and their variants are popular parameter-efficient fine-tuning (PEFT) techniques that closely match full model fine-tune performance while requiring only a…