Research map / Learning theory and algorithms / Prediction and decision losses
Invariant Subspace Decomposition
Margherita Lazzaretto, Jonas Peters, Niklas Pfister
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 |
|---|---|---|---|
| estimate invariant component with ISD instead of OLS on all training data (zero-shot task) | ISD invariant component · simulated time-varying regression | cumulative explained variance at new time points | Improved |
| estimate invariant component with ISD instead of maximin effect | ISD invariant component · simulated time-varying regression | explained variance under shifted test distribution | Improved |
| use ISD adaptation (invariant plus residual estimated on window) instead of OLS on rolling window | ISD time adaptation · simulation, adaptation window m=16 | MSPE difference OLS minus ISD | Improved |
| increase historical sample size n | estimated invariant component · simulation with non-smooth gamma variation | MSE of invariant component | Improved |
| increase adaptation window size m | ISD versus OLS adaptation · simulation MSPE comparison | MSPE gain of ISD over OLS | Hurt |
Closest papers by meaning
- PAC Generalization via Invariant Representations (ICML 2023 · similarity 0.86)
- Deconvolution in unlinked linear models (JMLR 2026 · similarity 0.86)
- The Implicit Bias of Heterogeneity towards Invariance: A Study of Multi-Environment Matrix Sensing (NeurIPS 2024 · similarity 0.86)
- Invariant Causal Representation Learning for Out-of-Distribution Generalization (ICLR 2022 · similarity 0.86)
- Demystifying Spectral Feature Learning for Instrumental Variable Regression (NeurIPS 2025 · similarity 0.85)
- Learning with Exact Invariances in Polynomial Time (ICML 2025 · similarity 0.84)
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.
