Research map / Probabilistic and scientific modeling / Bayesian optimization
Multi-fidelity hyperparameter optimization
Tune hyperparameters efficiently by allocating budget across fidelities and early stopping.
44 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2017–2026) · an approach in Bayesian optimization, keyword cluster “acquisition · bayesian · multi objective”
State History Tried Checklist Preprints Who Export
Where it stands
Not enough benchmark data Fewer than four papers report the same benchmark, so there is no record to follow.
10 papers in 2023–24 and 7 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 10.9% to 5.2%.
How this is decided
A benchmark is tracked when at least four papers of this approach report their own method on the same dataset and metric in at least two different years. Its record is the best number reported up to 2024 compared with the best reported in 2025–26: still improving means at least 1% better, flat means at most 0.2% better with at least two papers from 2025–26 trying. Numbers on a different scale from the rest (a fraction among percentages) and far outliers are left out. Results come from different papers with different setups, so they are signals, not head-to-head comparisons. The share of papers is counted at ICML, NeurIPS, ICLR, CVPR and AAAI.
How it got here
Most cited paper of each year
- 2017 · Learning Curve Prediction with Bayesian Neural Networks (ICLR 2017 · 170 citations)
- 2018 · BOHB: Robust and Efficient Hyperparameter Optimization at Scale (ICML 2018 · 180 citations)
- 2019 · Safe Exploration for Interactive Machine Learning (NeurIPS 2019 · 21 citations)
- 2020 · Efficient Automatic CASH via Rising Bandits (AAAI 2020 · 27 citations)
- 2021 · Warm Starting CMA-ES for Hyperparameter Optimization (AAAI 2021 · 35 citations)
- 2022 · Modeling the Machine Learning Multiverse (NeurIPS 2022 · 9 citations)
- 2023 · Human-in-the-Loop Optimization for Deep Stimulus Encoding in Visual Prostheses (NeurIPS 2023 · 5 citations)
- 2024 · PG-LBO: Enhancing High-Dimensional Bayesian Optimization with Pseudo-Label and Gaussian Process Guidance (AAAI 2024 · 7 citations)
- 2025 · Optimization Proxies using Limited Labeled Data and Training Time -- A Semi-Supervised Bayesian Neural Network Approach (ICML 2025 · 0 citations)
- 2026 · Discount Model Search for Quality Diversity Optimization in High-Dimensional Measure Spaces (ICLR 2026 · 0 citations)
Earliest papers here
- Learning Curve Prediction with Bayesian Neural Networks (ICLR 2017 · 170 citations)
- Efficient Hyperparameter Optimization for Deep Learning Algorithms Using Deterministic RBF Surrogates (AAAI 2017 · 130 citations)
- An Empirical Bayes Approach to Optimizing Machine Learning Algorithms (NeurIPS 2017)
What has been tried
37 papers here report 322 measured changes (a component added, removed or swapped, compared with a baseline): helped 66%, no clear effect 11%, hurt 15%, mixed 7%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Add pseudo-label training | Topology shape fitting · cosine similarity | Helped | PG-LBO: Enhancing High-Dimensional Bayesian Optimization with Pseudo-Label and Gaussian Process Guidance (AAAI 2024 · 7 citations) | |
| use adaptive efficient-point fidelity instead of a fixed fidelity | Slice · anytime validation loss | Helped | Efficient Hyperparameter Optimization with Adaptive Fidelity Identification (CVPR 2024 · 7 citations) | |
| adapt α rather than fix it to a constant | Eight MBO tasks, top-1 · average rank (lower is better) | fixed α ranks: 4.5, 3.9, 3.6, 2.4, 4.8; aSCR: 1.9 | Helped | Generative Adversarial Model-Based Optimization via Source Critic Regularization (NeurIPS 2024 · 2 citations) |
| optimize F1 and exact match jointly rather than F1 only | SQuAD2.0 · exact match | 75.09 -> 75.79 | Helped | Model Fusion through Bayesian Optimization in Language Model Fine-Tuning (NeurIPS 2024 · 1 citation) |
| Add Selection via Posterior | ACOPF case57 · Max Eq. | 0.045 -> 0.027 | Helped | Optimization Proxies using Limited Labeled Data and Training Time -- A Semi-Supervised Bayesian Neural Network Approach (ICML 2025 · 0 citations) |
| Add a time-dependent kernel to the GP | Offline historical data · cumulative regret | No clear effect | Multi-Timescale Dynamics Model Bayesian Optimization for Plasma Stabilization in Tokamaks (ICML 2025 · 0 citations) | |
| Use 32 shorter truncated runs instead of fewer, longer runs at the same FLOPS budget | Predicting final model losses · R² | 0.69 -> 0.82 | Helped | Data Mixture Optimization: A Multi-fidelity Multi-scale Bayesian Framework (NeurIPS 2025 · 0 citations) |
What reviewers will expect
Datasets and baselines named in the 10 papers of this approach from 2024–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Baselines
- Random search · 2 of 10 papers
- Hyperband · 2 of 10 papers
- BOHB · 2 of 10 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 17 papers; from OpenAlex and the papers' affiliations).
- Massachusetts Institute of Technology · 2 papers
Use it
Download BibTeX of 31 key papersOpen the topic in the interactive map
Get alerts. Describe your idea in scime atlas and save it (free scime account): you get a weekly email when new preprints land close to it.
Related approaches
Same keyword cluster
- Acquisition function design (97 papers)
- Local and high-dimensional Bayesian optimization (90 papers)
- Multi-objective Bayesian optimization (33 papers)
- Regret bounds for Bayesian optimization (43 papers)
Elsewhere in the topic
- Offline black-box optimization with surrogates (56 papers)
- High-dimensional Bayesian optimization (34 papers)
Benchmark numbers, ablations and baseline numbers are extracted by AI from each paper's full text and can be wrong; the paper is the authority. Data: results index of 2026-11-02, citation counts from OpenAlex (2026-10-09).