Inverse Problems Leveraging Pre-trained Contrastive Representations
6. prosince 2021
Řečníci
Sriram Ravula
Řečník · 0 sledujících
Georgios Smyrnis
Řečník · 0 sledujících
Matt Jordan
Řečník · 0 sledujících
Alex G. Dimakis
Řečník · 0 sledujících
O prezentaci
We study a new family of inverse problems for recovering representations of corrupted data. We assume access to a pre-trained representation learning network R(x) that operates on clean images, like CLIP. The problem is to recover the representation of an image R(x), if we are only given a corrupted version A(x), for some known forward operator A. We propose a supervised inversion method that uses a contrastive objective to obtain excellent representations for highly corrupted images. Using a linear probe on our robust representations, we achieve a higher accuracy than end-to-end supervised baselines when classifying images with various types of distortions, including blurring, additive noise, and random pixel masking. We evaluate on a subset of ImageNet and observe that our method is robust to varying levels of distortion. Our method outperforms end-to-end baselines even with a fraction of the labeled data in a wide range of forward operators.We study a new family of inverse problems for recovering representations of corrupted data. We assume access to a pre-trained representation learning network R(x) that operates on clean images, like CLIP. The problem is to recover the representation of an image R(x), if we are only given a corrupted version A(x), for some known forward operator A. We propose a supervised inversion method that uses a contrastive objective to obtain excellent representations for highly corrupted images. Using a li…
Organizátor
NeurIPS 2021
Účet · 1,9k sledujících
O organizátorovi (NeurIPS 2021)
Neural Information Processing Systems (NeurIPS) is a multi-track machine learning and computational neuroscience conference that includes invited talks, demonstrations, symposia and oral and poster presentations of refereed papers. Following the conference, there are workshops which provide a less formal setting.
Baví vás formát? Nechte SlidesLive zachytit svou akci!
Profesionální natáčení a streamování po celém světě.
Sdílení
Doporučená videa
Prezentace na podobné téma, kategorii nebo přednášejícího
OmniPrint: A Configurable Printed Character Synthesizer
Zhlédnout později
Haozhe Sun, …
Pro uložení prezentace do věčného trezoru hlasovalo 0 diváků, což je 0.0 %
Risk Bounds and Calibration for a Smart Predict-then-Optimize Method
Zhlédnout později
Heyuan Liu, …
Pro uložení prezentace do věčného trezoru hlasovalo 0 diváků, což je 0.0 %
Towards a Trace-Preserving Tensor Network Representation of Quantum Channels
Zhlédnout později
Pro uložení prezentace do věčného trezoru hlasovalo 0 diváků, což je 0.0 %
Offline Reinforcement Learning as One Big Sequence Modeling Problem
Zhlédnout později
Pro uložení prezentace do věčného trezoru hlasovalo 0 diváků, což je 0.0 %
The Consequences of Massive Scaling in Machine Learning
Zhlédnout později
Pro uložení prezentace do věčného trezoru hlasovalo 0 diváků, což je 0.0 %
Directed Graph Contrastive Learning
Zhlédnout později
Zekun Tong, …
Pro uložení prezentace do věčného trezoru hlasovalo 0 diváků, což je 0.0 %