Task Discovery: Finding the Tasks that Neural Networks Generalize on
Dec 6, 2022
Speakers
Andrei Atanov
Speaker · 0 followers
Andrei Filatov
Speaker · 0 followers
Teresa Yeo
Speaker · 0 followers
Ajay Sohmshetty
Speaker · 0 followers
Amir Zamir
Speaker · 1 follower
About
When developing deep learning models, we usually decide what task we want to solve then search in the space of models in order to design one that generalizes well on this task. An intriguing question would be: what if, instead of fixing the task and searching in the model space, we fix the model and search in the task space? Can we find tasks that the model generalizes on? What do they look like, or do they show anything?This is the question we address in this paper. We propose a task discovery framework that automatically finds examples of such tasks via optimizing a generalization-based quantity called agreement score. With this framework, we demonstrate that the same set of images can allow for many tasks on which neural networks generalize well. The understandings from task discovery can also provide a tool to shed more light on deep learning and its failure modes: as an example, we show that the discovered tasks can be used to generate “adversarial train-test splits" which make a model fail at test time, without changing the pixels or labels, but only by selecting how the datapoints should be split between training and testing.When developing deep learning models, we usually decide what task we want to solve then search in the space of models in order to design one that generalizes well on this task. An intriguing question would be: what if, instead of fixing the task and searching in the model space, we fix the model and search in the task space? Can we find tasks that the model generalizes on? What do they look like, or do they show anything?This is the question we address in this paper. We propose a task discovery…
Organizer
NeurIPS 2022
Account · 961 followers
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
Identify event causality with knowledge and analogy
Watch later
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Fast Mixing of Stochastic Gradient Descent with Normalization and Weight Decay
Watch later
Zhiyuan Li, …
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Personalizing Sustainable Agriculture with Causal Machine Learning
Watch later
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
FinRL-Meta: Market Environments and Benchmarks for Data-Driven Financial Reinforcement Learning
Watch later
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Generating Long Videos of Dynamic Scenes
Watch later
Tim Brooks, …
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
C-Mixup: Improving Generalization in Regression
Watch later
Huaxiu Yao, …
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%