Research map / Probabilistic and scientific modeling / Treatment effects
Optimized Covariance Design for AB Test on Social Network under Interference
Qianyi Chen, Bo Li, LU DENG, Yong Wang
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 optimized covariance design (OCD) instead of complete randomization | cluster randomization with HT estimator · FB-Stanford3 multiplicative model, gamma=2.0 | MSE: 2.297 -> 1.073 | Improved |
| use OCD instead of PSR | cluster randomization with HT estimator · FB-Stanford3 linear model, gamma=2.0 | MSE: 1.496 -> 0.926 | Improved |
| use OCD instead of IBR-p at weak interference | cluster randomization with HT estimator · FB-Stanford3 linear model, gamma=0.5 | MSE: 0.141 -> 0.209 | Hurt |
| use OCD instead of PSR at gamma=1.0 | cluster randomization with HT estimator · FB-Stanford3 linear model, gamma=1.0 | MSE: 0.415 -> 0.371 | Improved |
| use OCD instead of PSR | cluster randomization with HT estimator · FB-Stanford3 multiplicative model, gamma=0.5 | MSE: 0.152 -> 0.069 | Improved |
| use OCD instead of complete randomization | cluster randomization with HT estimator · FB-Stanford3 linear model, gamma=2.0 | SD: 0.562 -> 0.585 | No clear effect |
| use OCD instead of other designs under model misspecification (multiplicative model) | cluster randomization with HT estimator · FB-Stanford3 multiplicative model, gamma=1.0 | MSE: 0.569 -> 0.271 | Improved |
| vary Louvain clustering resolution (2, 5, 10) (ablation) | OCD · FB-Stanford3 and FB-Cornell5 appendix | MSE | No clear effect |
| vary omega in the MSE bound (ablation) | OCD · FB-Stanford3 | MSE | No clear effect |
| use weaker spectral-norm bound on interference term (ablation) | OCD objective · FB-Stanford3 | MSE | Hurt |
| use worst-case variance objective (IBR) instead of OCD bound | block randomization · FB-Stanford3 | MSE: 0.926 -> 1.541 | Hurt |
| use pairing blocks of size 2 (IBR-p) instead of IBR | independent block randomization · FB-Stanford3 linear model, gamma=0.5 | MSE: 0.164 -> 0.141 | Improved |
Closest papers by meaning
- Just Ramp-Up: Debiasing Regression-based Estimator for A/B Tests under Network Interference (NeurIPS 2026 · similarity 0.90)
- Online Experimental Design With Estimation-Regret Trade-off Under Network Interference (NeurIPS 2025 · similarity 0.88)
- Independent-Set Design of Experiments for Estimating Treatment and Spillover Effects under Network Interference (ICLR 2024 · similarity 0.86)
- Doubly Robust Causal Effect Estimation under Networked Interference via Targeted Learning (ICML 2024 · similarity 0.85)
- Higher-Order Causal Message Passing for Experimentation with Complex Interference (NeurIPS 2024 · similarity 0.85)
- Active Treatment Effect Estimation via Limited Samples (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.
