Research map / Probabilistic and scientific modeling / Bayesian optimization
Task-Agnostic Amortized Multi-Objective Optimization
Xinyu Zhang, Conor Hassan, Julien Martinelli, Daolang Huang, Samuel Kaski
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 |
|---|---|---|---|
| add prediction warm-up and prediction loss term to policy training (ablation) | TAMO · synthetic MOO tasks | simple regret | Improved |
| increase prediction loss weight lambda_p (0.25 to 1.0) (ablation) | TAMO · synthetic MOO tasks | simple regret | No clear effect |
| vary discount factor gamma (ablation) | TAMO · synthetic MOO tasks | simple regret | No clear effect |
| train with full 100-step horizon instead of myopic horizon T=1 (ablation) | TAMO · synthetic MOO tasks | simple regret | Mixed |
| use 4 transformer layers per module instead of 2 (ablation) | TAMO · synthetic MOO tasks | simple regret | Improved |
| use pretraining prior with smaller lengthscales (ablation) | TAMO · synthetic MOO tasks | simple regret | Hurt |
| add quadratic-bowl term to pretraining prior (ablation) | TAMO · synthetic MOO tasks | simple regret | Mixed |
| increase batch size q from 1 to 10 (ablation) | TAMO · synthetic MOO tasks | simple regret: q in {1, 2, 5, 10} | Hurt |
| increase query set size Nq from 256 to 2048 (ablation) | TAMO · synthetic MOO tasks | simple regret: Nq in {256, 512, 1024, 2048} | No clear effect |
| use decoupled objective observations | TAMO · synthetic MOO tasks | regret vs cost | Mixed |
Closest papers by meaning
- Preference-Aware Constrained Multi-Objective Bayesian Optimization (Student Abstract) (AAAI 2024 · similarity 0.88)
- SMOG: Scalable Meta-Learning for Multi-Objective Bayesian Optimization (NeurIPS 2026 · similarity 0.88)
- Optimizing the Unknown: Black Box Bayesian Optimization with Energy-Based Model and Reinforcement Learning (NeurIPS 2025 · similarity 0.87)
- BOtied: Multi-objective Bayesian optimization with tied multivariate ranks (ICML 2024 · similarity 0.87)
- Simple Extensions of Single-Objective Acquisition Functions and Hedge Strategies for Multi-Objective Bayesian Optimization (NeurIPS 2026 · similarity 0.87)
- Pareto Front-Diverse Batch Multi-Objective Bayesian Optimization (AAAI 2024 · similarity 0.87)
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.
