Research map / Probabilistic and scientific modeling / Variational and generative sampling
Offline Model-Based Optimization via Normalized Maximum Likelihood Estimation
Justin Fu, Sergey Levine
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 NEMO's quantized NML training instead of optimizing a pretrained forward model without NML (ablation) | NEMO · Superconductor, MoleculeActivity, and GFP | ground truth score | Improved |
| replace the discretized logistic architecture with a categorical feedforward network (ablation) | NEMO · Superconductor | 50th-percentile ground truth score | Hurt |
| replace the discretized logistic architecture with a categorical feedforward network (ablation) | NEMO · MoleculeActivity | ground truth score | Hurt |
| increase model optimization steps while using a small model learning rate (ablation) | NEMO · Superconductor | ground truth score | Improved |
| increase model optimization steps while using a high model learning rate (ablation) | NEMO · Superconductor | ground truth score | Hurt |
| increase model optimization steps while using a small model learning rate (ablation) | NEMO · MoleculeActivity | ground truth score | Improved |
| increase model optimization steps while using a high model learning rate (ablation) | NEMO · MoleculeActivity | ground truth score | No clear effect |
| use NEMO instead of the strongest baseline | offline MBO · 100th-percentile scores across six Design-bench tasks | ground truth score | Mixed |
| use NEMO instead of the strongest baseline | offline MBO · 50th-percentile scores across six Design-bench tasks | ground truth score | Mixed |
Closest papers by meaning
- ROOT: Rethinking Offline Optimization as Distributional Translation via Probabilistic Bridge (NeurIPS 2025 · similarity 0.87)
- A Study of Bayesian Neural Network Surrogates for Bayesian Optimization (ICLR 2024 · similarity 0.87)
- Adaptive Partitioning Schemes for Optimistic Optimization (ICML 2025 · similarity 0.86)
- Learning Surrogates for Offline Black-Box Optimization via Gradient Matching (ICML 2024 · similarity 0.86)
- Diversity By Design: Leveraging Distribution Matching for Offline Model-Based Optimization (ICML 2025 · similarity 0.86)
- Offline Model-Based Optimization via Policy-Guided Gradient Search (AAAI 2024 · similarity 0.86)
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.
