Research map / Probabilistic and scientific modeling / Bayesian optimization
Adaptive Acquisition Selection for Bayesian Optimization with Large Language Models
Giang Ngo, Dat Phan Trong, Dang Nguyen, Sunil Gupta, Svetha Venkatesh
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 remaining budget information to the LLM prompt (ablation) | LMABO · 50 BO problems | Mean RP: 1.40 -> 1.21 | Improved |
| add GP model characteristics to the prompt (ablation) | LMABO · 50 BO problems | Mean RP: 1.50 -> 1.21 | Improved |
| add shortest distance information to the prompt (ablation) | LMABO · 50 BO problems | Mean RP: 1.50 -> 1.21 | Improved |
| instruct the LLM to avoid AFs that failed to improve the incumbent (ablation) | LMABO · 50 BO problems | Mean RP: 1.92 -> 1.21 | Improved |
| use Qwen3-8B instead of Gemini-2.5 Flash (ablation) | LMABO · 50 BO problems | Mean RP: 1.21 -> 1.48 | Hurt |
| use Qwen3-30B thinking model instead of Gemini-2.5 Flash (ablation) | LMABO · 50 BO problems | Mean RP: 1.21 -> 1.29 | Hurt |
| use gpt-oss-120b or GPT-4o mini instead of Gemini-2.5 Flash (ablation) | LMABO · 50 BO problems | Mean RP: 1.21 -> 1.22 (120B), 1.21 (GPT-4o mini) | No clear effect |
| use LLM-chosen acquisition function over best static AF (EI) | best static AF · 50 BO problems | Mean RP: 1.34 -> 1.21 | Improved |
| use LLM-chosen AF selection over LLM-based BO baselines | LLAMBO / LLMP · 50 BO problems | Mean RP: 2.67 -> 1.21 | Improved |
| restrict adaptive portfolio baselines to curated EI, LogEI, TS set (ablation) | adaptive portfolio methods · 50 BO problems | AUC: no numbers in text | Mixed |
| add task-specific context to initial prompt (ablation) | LMABO · 4 benchmark problems | regret curves on 4 problems | Improved |
Closest papers by meaning
- Large Language Models to Enhance Bayesian Optimization (ICLR 2024 · similarity 0.90)
- Adaptive Kernel Design for Bayesian Optimization Is a Piece of CAKE with LLMs (NeurIPS 2025 · similarity 0.88)
- Simple Extensions of Single-Objective Acquisition Functions and Hedge Strategies for Multi-Objective Bayesian Optimization (NeurIPS 2026 · similarity 0.87)
- Tree-Structured Synergy of Large Language Models and Bayesian Optimization for Efficient CASH (NeurIPS 2026 · similarity 0.87)
- FunBO: Discovering Acquisition Functions for Bayesian Optimization with FunSearch (ICML 2025 · similarity 0.87)
- Batched Energy-Entropy acquisition for Bayesian Optimization (NeurIPS 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.
