On Interaction Between Augmentations and Corruptions in Natural Corruption Robustness
Dec 6, 2021
Speakers
Eric Mintun
Speaker · 0 followers
Alexander Kirillov
Speaker · 0 followers
Saining Xie
Speaker · 0 followers
About
Invariance to a broad array of image corruptions, such as warping, noise, or color shifts, is an important aspect of building robust models in computer vision. Recently, several new data augmentations have been proposed that significantly improve performance on ImageNet-C, a benchmark of such corruptions. However, there is still a lack of basic understanding on the relationship between data augmentations and test-time corruptions. To this end, we develop a feature space for image transforms, and then use a new measure in this space between augmentations and corruptions called the Minimal Sample Distance to demonstrate there is a strong correlation between similarity and performance. We then investigate recent data augmentations and observe a significant degradation in corruption robustness when the test-time corruptions are sampled to be perceptually dissimilar from ImageNet-C in this feature space. Our results suggest that test error can be improved by training on perceptually similar augmentations, and data augmentations may not generalize well beyond the existing benchmark. We hope our results and tools will allow for more robust progress towards improving robustness to image corruptions.Invariance to a broad array of image corruptions, such as warping, noise, or color shifts, is an important aspect of building robust models in computer vision. Recently, several new data augmentations have been proposed that significantly improve performance on ImageNet-C, a benchmark of such corruptions. However, there is still a lack of basic understanding on the relationship between data augmentations and test-time corruptions. To this end, we develop a feature space for image transforms, and…
Organizer
NeurIPS 2021
Account · 1.9k followers
About 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.
Like the format? Trust SlidesLive to capture your next event!
Professional recording and live streaming, delivered globally.
Sharing
Recommended Videos
Presentations on similar topic, category or speaker
Multi-Agent Reinforcement Learning in Stochastic Networked Systems
Watch later
Yiheng Lin, …
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
REAL2: An end-to-end memory-augmented solver for math word problems
Watch later
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Discussion Panel 1: Decision Making
Watch later
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Meta-Learning an Inference Algorithm for Probalistic Programs
Watch later
Gwonsoo Che, …
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Performance Analysis of Quantum Machine Learning Classifiers
Watch later
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Panel Discussion 2
Watch later
Moritz Hardt, …
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%