Research map / Label-efficient and robust learning / Domain adaptation and generalization
Learning Conditional Invariances through Non-Commutativity
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Proposes Non-Commutative Invariance (NCI), relaxing invariance learning to be directed toward a target domain so conditional invariances can be learned provably optimally and sample-efficiently under domain asymmetry.
Problem
Invariance learning methods filter out domain-specific distractor variables based only on data semantics, without regard to the target domain under evaluation, which limits generalization under domain shift.
Approach
Relax the invariance criterion to be non-commutatively directed toward the target domain, so that conditional invariances are learned as an asymmetric operation from source to target, with theoretical optimality and sample-efficiency guarantees under domain asymmetry.
Key results
- Evaluates NCI against DANN and CICyc (Samarin et al., 2021) on PACS, Office-Home, and DomainNet, reported in Tables 1 and 4 of the rebuttal version.
Limitations
- Text provided is only a rebuttal, so the full method and results are not fully visible.
- Theoretical guarantees depend on an asymmetry assumption between domains.
Experimental setup
Datasets: PACS Office-Home DomainNet
Benchmarks: PACS Office-Home DomainNet
Metrics: accuracy
Baselines: DANN CICyc
Builds on: DANN domain-adversarial training H-divergence
Closest papers by meaning
- Learning Invariances using the Marginal Likelihood (NeurIPS 2018 · similarity 0.87)
- The Implicit Bias of Heterogeneity towards Invariance: A Study of Multi-Environment Matrix Sensing (NeurIPS 2024 · similarity 0.85)
- Continuous Invariance Learning (ICLR 2024 · similarity 0.85)
- Learning with Exact Invariances in Polynomial Time (ICML 2025 · similarity 0.85)
- Guarantees for Nonlinear Representation Learning: Non-identical Covariates, Dependent Data, Fewer Samples (ICML 2024 · similarity 0.85)
- Robust Domain Generalization under Divergent Marginal and Conditional Distributions (NeurIPS 2026 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.