Inversion of Deep Generative Models

Time：Mon., 16:00 March 3, 2025

Venue：Voov (Tencent): 532-632-685

Organizer：Angelica Aviles-Rivero

Speaker：Seongmin Hong

Math+ML+X Seminar Series

Organizer：

Angelica Aviles-Rivero

Speaker：

Seongmin Hong (Seoul National University)

Time：

Mon., 16:00

March 3, 2025

Online：

Voov (Tencent): 532-632-685

Title:

Inversion of Deep Generative Models

Abstract：

In this seminar, Seongmin Hong will present his recent work on deep generative models, focusing on inversion methods. Topics will include:

--Gradient-free decoder inversion in latent diffusion models

--Exact inversion techniques in DPM-solvers

--The robustness of normalizing flows for inverse problems

--Neural diffeomorphic non-uniform B-spline flows

DATEMarch 2, 2025

Related News

0
Deep image prior for inverse problems: acceleration and probabilistic treatment
AbstractSince its first proposal in 2018, deep image prior has emerged as a very powerful unsupervised deep learning technique for solving inverse problems. The approach has demonstrated very encouraging empirical success in image denoising, deblurring, super-resolution etc. However, there are also several known drawbacks of the approach, notably high computational expense. In this talk, we des...
1
Algebraic Geometry Seminar | Adjunction and inversion of adjunction
AbstractIn this talk, I will explain adjunction and inversion of adjunction for log canonical centers of arbitrary codimension. Our result depends on the existence of log canonical modifications and the theory of basic slc-trivial fibrations. Hence it depends on the minimal model program for log canonical pairs and the theory of variations of mixed Hodge structure. This talk is based on a joint...