Ben Adcock, Deep learning for inverse problems: confident hallucinations and new theory 2026.02.03
Автор: CodEx Seminar
Загружено: 2026-02-03
Просмотров: 102
Описание:
Speaker: Ben Adcock (Simon Fraser University)
Title: Deep learning for inverse problems: confident hallucinations and new theoretical guarantees for Bayesian recovery.
Date: February 03, 2026
Abstract: Deep learning is currently transforming how inverse problems arising in imaging reconstruction are solved. However, it is increasingly well-known that such deep learning-based methods are susceptible to hallucinations. In this talk, I will present several theoretical explanations for why hallucinations occur, in both deterministic and statistical estimators. I will conclude by observing that hallucinations can only be avoided by careful design of the forwards operator in tandem with the recovery algorithm, and then provide a theoretical framework for how this can be achieved in a Bayesian setting in the case of posterior sampling using generative models.
Notes:
Video title edited to fit length requirements.
Video lightly edited at 2:40 on account of technical difficulties
Повторяем попытку...
Доступные форматы для скачивания:
Скачать видео
-
Информация по загрузке:
