PhD student, Massachusetts Institute of Technology
1 paper at NeurIPS 2025
We improve the quality of generative models by using low-quality, corrupted, and out-of-distribution data
