17 citations · 22 across the 5 of their papers we have counts for
8 papers
GIVT: Generative Infinite-Vocabulary Transformers
Michael Tschannen, Cian Eastwood, Fabian Mentzer
We introduce Generative Infinite-Vocabulary Transformers (GIVT) which generate vector sequences with real-valued entries, instead of discrete tokens from a finite vocabulary. To th…
Self-Supervised Disentanglement by Leveraging Structure in Data Augmentations
Cian Eastwood, Julius von Kügelgen, Linus Ericsson +4
Self-supervised representation learning often uses data augmentations to induce some invariance to "style" attributes of the data. However, with downstream tasks generally unknown…
Spuriosity Didn't Kill the Classifier: Using Invariant Predictions to Harness Spurious Features
Cian Eastwood, Shashank Singh, Andrei Liviu Nicolicioiu +3
To avoid failures on out-of-distribution data, recent works have sought to extract features that have an invariant or stable relationship with the label across domains, discarding…
DCI-ES: An Extended Disentanglement Framework with Connections to Identifiability
Cian Eastwood, Andrei Liviu Nicolicioiu, Julius von Kügelgen +4
In representation learning, a common approach is to seek representations which disentangle the underlying factors of variation. Eastwood & Williams (2018) proposed three metrics fo…
Probable Domain Generalization via Quantile Risk Minimization
Cian Eastwood, Alexander Robey, Shashank Singh +4
Domain generalization (DG) seeks predictors which perform well on unseen test distributions by leveraging data drawn from multiple related training distributions or domains. To ach…
Align-Deform-Subtract: An Interventional Framework for Explaining Object Differences
Cian Eastwood, Li Nanbo, Christopher K. I. Williams
Given two object images, how can we explain their differences in terms of the underlying object properties? To address this question, we propose Align-Deform-Subtract (ADS) -- an i…