Research map / Learning theory and algorithms / Submodular and game algorithms
Elementary Symmetric Polynomials for Optimal Experimental Design
Zelda E. Mariet, Suvrit Sra
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 greedy removal initialized from the continuous relaxation's support instead of sampling from the relaxed solution | ESP-design · Synthetic experiments | design objective quality | Improved |
| Use greedy removal rather than uniform initialization with Fedorov exchange | ESP-design · Synthetic experiments | design objective quality | No clear effect |
| Run Fedorov exchange after greedy initialization | ESP-design · Synthetic experiments | design objective quality | Improved |
| Increase the ESP order ℓ | ESP-design · Concrete Compressive Strength | fraction of non-zero entries in the design matrix | Improved |
| Increase the ESP order ℓ | ESP-design · Concrete Compressive Strength | normalized prediction error | Hurt |
| Use A-optimal design instead of higher-order ESP design | ESP-design · Concrete Compressive Strength | normalized prediction error | Improved |
| Initialize Fedorov exchange with the greedy design rather than uniform sampling | ESP-design · Synthetic experiments, ℓ=10, d=0.6 | runtime (s): 1.6 $ 10^3$ vs 6.6 $ 10^1$ at k=40 | Improved |
Closest papers by meaning
- Near-Optimal Design of Experiments via Regret Minimization (ICML 2017 · similarity 0.85)
- Polynomial Preconditioning for Gradient Methods (ICML 2023 · similarity 0.83)
- Batch greedy maximization of non-submodular functions: Guarantees and applications to experimental design (JMLR 2021 · similarity 0.82)
- Near-Optimal Experiment Design in Linear non-Gaussian Cyclic Models (NeurIPS 2025 · similarity 0.82)
- Finding Simple Proofs for First-Order Optimization (NeurIPS 2026 · similarity 0.82)
- The Non-linear $F$-Design and Applications to Interactive Learning (ICML 2024 · similarity 0.81)
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.
