AutoML in Heavily Constrained Applications
Not on the map. Why?
Abstract (arXiv)
Optimizing a machine learning pipeline for a task at hand requires careful configuration of various hyperparameters, typically supported by an AutoML system that optimizes the hyperparameters for the given training dataset. Yet, depending on the AutoML system's own second-order meta-configuration, the performance of the AutoML process can vary significantly. Current AutoML systems cannot automatically adapt their own configuration to a specific use case. Further, they cannot compile user-defined application constraints on the effectiveness and efficiency of the pipeline and its generation. In this paper, we propose CAML, which uses meta-learning to automatically adapt its own AutoML parameters, such as the search strategy, the validation strategy, and the search space, for a task at hand. The dynamic AutoML strategy of CAML takes user-defined constraints into account and obtains constraint-satisfying pipelines with high predictive performance.
Key facts. Cited by 3 papers in the map (data through 7 Oct 2026). Builds on 29 of its 78 references that are in the map. First posted on arXiv 29 Jun 2023.
Builds on
29 of 78 references.
References · 29
- 2019Optuna: A Next-generation Hyperparameter Optimization Frameworkcited by 2,455
- 2016Equality of Opportunity in Supervised Learningcited by 2,000
- 2019Federated Learning: Challenges, Methods, and Future Directionscited by 1,531
- 2016Hyperband: A Novel Bandit-Based Approach to Hyperparameter Optimizationcited by 688
- 2014OpenML: networked science in machine learningcited by 523
- 2018BOHB: Robust and Efficient Hyperparameter Optimization at Scalecited by 466
- 2021Retiring Adult: New Datasets for Fair Machine Learningcited by 443
- 2020AutoGluon-Tabular: Robust and Accurate AutoML for Structured Datacited by 368
- 2019Machine Learning Testing: Survey, Landscapes and Horizonscited by 213
- 2020Auto-Sklearn 2.0: Hands-free AutoML via Meta-Learningcited by 174
- 2014Bayesian Optimization with Unknown Constraintscited by 109
- 2017Hyperparameter Importance Across Datasetscited by 62
17 more
- 2019An ADMM Based Framework for AutoML Pipeline Configurationcited by 47
- 2020Fair Bayesian Optimizationcited by 45
- 2021VolcanoML: Speeding up End-to-End AutoML via Scalable Search Space Decompositioncited by 25
- 2021Towards Green Automated Machine Learning: Status Quo and Future Directionscited by 25
- 2021Emulation of physical processes with Emukitcited by 24
- 2019Overton: A Data System for Monitoring and Improving Machine-Learned Productscited by 21
- 2019SystemDS: A Declarative Machine Learning System for the End-to-End Data Science Lifecyclecited by 19
- 2017GPflowOpt: A Bayesian Optimization Library using TensorFlowcited by 18
- 2019Towards Assessing the Impact of Bayesian Optimization's Own Hyperparameterscited by 16
- 2016KeystoneML: Optimizing Pipelines for Large-Scale Advanced Analyticscited by 13
- 2019FairPrep: Promoting Data to a First-Class Citizen in Studies on Fairness-Enhancing Interventionscited by 13
- 2018Helix: Holistic Optimization for Accelerating Iterative Machine Learningcited by 11
- 2021Auto-Pipeline: Synthesizing Complex Data Pipelines By-Target Using Reinforcement Learning and Searchcited by 8
- 2017Real-Time Machine Learning: The Missing Piecescited by 7
- 2022DiffML: End-to-end Differentiable ML Pipelinescited by 6
- 2021An Efficient Asynchronous Batch Bayesian Optimization Approach for Analog Circuit Synthesiscited by 5
- 2017A Cost-based Optimizer for Gradient Descent Optimizationcited by 4
Built on by
Citing papers per year: 2024: 1; 2025: 2.
3 papers.