HOrIc&RRO7HrEs#uNl NB%6GDjxO2Y0

My name is Justin Deschenaux and I am a PhD student at . I am advised by Professor Caglar Gulcehre, and I work on diffusion language models, a blazingly fast, controllable, and principled way to generate text. My research focuses on (1) fast inference through distillation and efficient architectures (SDTT, PGM, BlockGen); (2) bridging continuous and discrete diffusion to transfer powerful techniques (Diffusion Duality Chapter I, Chapter II); and (3) samplers that improve generation quality (Ψ-Samplers, Loopholing, BlockGen).