Research map / Probabilistic and scientific modeling / Molecules and proteins
Homogeneity Structure Learning in Large-scale Panel Data with Heavy-tailed Errors
Di Xiao, Yuan Ke, Runze Li
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 robust (Huber) covariance estimation instead of sample covariance | latent factor estimation · Covariance estimation, Pareto errors | covariance error (max norm): 6.662 -> 0.297 | Improved |
| use robust covariance instead of sample covariance | latent factor estimation · Covariance estimation, t2.1 errors | covariance error (Frobenius): 0.917 -> 0.437 | Improved |
| use robust covariance instead of sample covariance | latent factor estimation · Covariance estimation, Normal errors | covariance error (max norm): 0.022 -> 0.022 | No clear effect |
| use adaptive Huber instead of median-of-means covariance | covariance estimation · Covariance estimation, Normal errors | covariance error (max norm): 0.027 -> 0.022 | Improved |
| compare fast data-driven robustification vs Ah | covariance estimation · Covariance estimation, Pareto errors | covariance error (max norm): 0.298 -> 0.297 | No clear effect |
| add robust estimation (vs Homogeneity estimator, OLS-based) (ablation) | homogeneity estimator · Coefficient estimation, heavy-tailed errors | RMSE | Improved |
| add homogeneity detection (vs Robust estimator without it) (ablation) | robust estimator · Coefficient estimation RMSE | RMSE | Improved |
| increase signal strength r from 1 to 4 | homogeneity detection · 5-groups, Pareto errors | adjusted Rand index: 0.9967 -> 0.9997 | Improved |
| increase signal strength r from 1 to 4 | homogeneity detection · 9-groups, Pareto errors | adjusted Rand index: 0.9940 -> 0.9998 | Improved |
| make errors serially dependent | homogeneity detection · 5-groups, r=1, Pareto errors | adjusted Rand index: 0.9967 -> 0.9865 | Hurt |
| vary robustification parameter tau (ablation) | robust covariance and robust regression · Sensitivity (Figure 11) | Frobenius error / RMSE | No clear effect |
| learn homogeneity structure with robust estimation instead of OLS | PM2.5 prediction · EPA air pollutant data, rolling window | out-of-sample MSE | Improved |
Closest papers by meaning
- Covariate-Adjusted Deep Causal Learning for Heterogeneous Panel Data Models (NeurIPS 2026 · similarity 0.85)
- A Meta-learner for Heterogeneous Effects in Difference-in-Differences (ICML 2025 · similarity 0.83)
- Scores for Learning Discrete Causal Graphs with Unobserved Confounders (AAAI 2024 · similarity 0.83)
- Local Learning for Covariate Selection in Nonparametric Causal Effect Estimation with Latent Variables (NeurIPS 2025 · similarity 0.82)
- Identification of Causal Structure with Latent Variables Based on Higher Order Cumulants (AAAI 2024 · similarity 0.82)
- IncomeSCM: From tabular data set to time-series simulator and causal estimation benchmark (NeurIPS 2024 · 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.
