Timezone: »
In the domain generalization literature, a common objective is to learn representations independent of the domain after conditioning on the class label. We show that this objective is not sufficient: there exist counter-examples where a model fails to generalize to unseen domains even after satisfying class-conditional domain invariance. We formalize this observation through a structural causal model and show the importance of modeling within-class variations for generalization. Specifically, classes contain objects that characterize specific causal features, and domains can be interpreted as interventions on these objects that change non-causal features. We highlight an alternative condition: inputs across domains should have the same representation if they are derived from the same object. Based on this objective, we propose matching-based algorithms when base objects are observed (e.g., through data augmentation) and approximate the objective when objects are not observed (MatchDG). Our simple matching-based algorithms are competitive to prior work on out-of-domain accuracy for rotated MNIST, Fashion-MNIST, PACS, and Chest-Xray datasets. Our method MatchDG also recovers ground-truth object matches: on MNIST and Fashion-MNIST, top-10 matches from MatchDG have over 50% overlap with ground-truth matches.
Author Information
Divyat Mahajan (Microsoft Research India)
Shruti Tople (Microsoft Research)
Amit Sharma (Microsoft Research)
Related Events (a corresponding poster, oral, or spotlight)
-
2021 Poster: Domain Generalization using Causal Matching »
Thu. Jul 22nd 04:00 -- 06:00 PM Room Virtual
More from the Same Authors
-
2021 : DoWhy: Addressing Challenges in Expressing and Validating Causal Assumptions »
Amit Sharma · Vasilis Syrgkanis · cheng zhang · Emre Kiciman -
2022 : Modeling the Data-Generating Process is Necessary for Out-of-Distribution Generalization »
JIVAT NEET KAUR · Emre Kiciman · Amit Sharma -
2022 : Probing Classifiers are Unreliable for Concept Removal and Detection »
Abhinav Kumar · Chenhao Tan · Amit Sharma -
2023 : Towards Modular Machine Learning Pipelines »
Aditya Modi · JIVAT NEET KAUR · Maggie Makar · Pavan Mallapragada · Amit Sharma · Emre Kiciman · Adith Swaminathan -
2023 Poster: Bayesian Estimation of Differential Privacy »
Santiago Zanella-Beguelin · Lukas Wutschitz · Shruti Tople · Ahmed Salem · Victor Ruehle · Andrew Paverd · Mohammad Naseri · Boris Köpf · Dan Jones -
2023 Poster: Synergies between Disentanglement and Sparsity: Generalization and Identifiability in Multi-Task Learning »
Sébastien Lachapelle · Tristan Deleu · Divyat Mahajan · Ioannis Mitliagkas · Yoshua Bengio · Simon Lacoste-Julien · Quentin Bertrand -
2023 Oral: Interventional Causal Representation Learning »
Kartik Ahuja · Divyat Mahajan · Yixin Wang · Yoshua Bengio -
2023 Poster: Interventional Causal Representation Learning »
Kartik Ahuja · Divyat Mahajan · Yixin Wang · Yoshua Bengio -
2022 Poster: Matching Learned Causal Effects of Neural Networks with Domain Priors »
Sai Srinivas Kancheti · Gowtham Reddy Abbavaram · Vineeth N Balasubramanian · Amit Sharma -
2022 Spotlight: Matching Learned Causal Effects of Neural Networks with Domain Priors »
Sai Srinivas Kancheti · Gowtham Reddy Abbavaram · Vineeth N Balasubramanian · Amit Sharma -
2021 Poster: Grey-box Extraction of Natural Language Models »
Santiago Zanella-Beguelin · Shruti Tople · Andrew Paverd · Boris Köpf -
2021 Spotlight: Grey-box Extraction of Natural Language Models »
Santiago Zanella-Beguelin · Shruti Tople · Andrew Paverd · Boris Köpf -
2020 Poster: Alleviating Privacy Attacks via Causal Learning »
Shruti Tople · Amit Sharma · Aditya Nori