Research map / Label-efficient and robust learning / Recognition and long-tail
Fundamental Limits and Tradeoffs in Invariant Representation Learning
Han Zhao, Chen Dan, Bryon Aragam, Tommi Jaakkola, Geoffrey Gordon, Pradeep Ravikumar
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use an MLP representation instead of Logit | UCI Adult · UCI Adult | I(Y;Z): 0.170 -> 0.216 | Improved |
| Use an MLP representation instead of Logit | UCI Adult · UCI Adult | I(A;Z): 0.103 -> 0.092 | Improved |
| Use an MLP representation instead of Linear | UCI Adult · UCI Adult | I(Y;Z): 0.171 -> 0.216 | Improved |
| Use an MLP representation instead of Linear | UCI Adult · UCI Adult | I(A;Z): 0.109 -> 0.092 | Improved |
| Use Adv instead of MLP | UCI Adult · UCI Adult | I(Y;Z): 0.216 -> 0.202 | Hurt |
| Use Adv instead of MLP | UCI Adult · UCI Adult | I(A;Z): 0.092 -> 0.068 | Improved |
| Use MLP instead of OLS | Law School · Law School | Var E[Y|Z]: 0.014 -> 0.022 | Improved |
| Use MLP instead of OLS | Law School · Law School | Var E[A|Z]: 0.008 -> 0.012 | Hurt |
| Use Adv instead of OLS | Law School · Law School | Var E[Y|Z]: 0.014 -> 0.018 | Improved |
| Use Adv instead of OLS | Law School · Law School | Var E[A|Z]: 0.008 -> 0.010 | Hurt |
| Randomize between existing representations | Invariant representation learning · Classification and regression information planes | Achievable accuracy-invariance tradeoffs | Improved |
Closest papers by meaning
- Learning Invariant Representations with a Nonparametric Nadaraya-Watson Head (NeurIPS 2023 · similarity 0.86)
- Invariant Representations without Adversarial Training (NeurIPS 2018 · similarity 0.86)
- Learning Invariant Representations using Inverse Contrastive Loss (AAAI 2021 · similarity 0.85)
- Controllable Invariance through Adversarial Feature Learning (NeurIPS 2017 · similarity 0.85)
- Max-Margin Invariant Features from Transformed Unlabelled Data (NeurIPS 2017 · similarity 0.84)
- On Making $SE(2)$-Invariant Networks Optimal (NeurIPS 2026 · similarity 0.82)
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.
