Model Selection
Model selection aims to identify the optimal machine learning model for a given task from a potentially vast pool of candidates, balancing performance with resource efficiency. Current research emphasizes efficient algorithms for selecting models, including those leveraging transfer learning, ensembling techniques (like top-k union), and meta-learning approaches to adapt to diverse data distributions and changing objectives. Effective model selection is crucial for improving the accuracy and reliability of machine learning applications across various domains, from natural language processing and computer vision to resource-constrained edge devices and cybersecurity, while also mitigating the environmental impact of computationally intensive model training and deployment.
Papers
Computation-Aware Gaussian Processes: Model Selection And Linear-Time Inference
Jonathan Wenger, Kaiwen Wu, Philipp Hennig, Jacob R. Gardner, Geoff Pleiss, John P. Cunningham
Dirichlet process mixtures of block $g$ priors for model selection and prediction in linear models
Anupreet Porwal, Abel Rodriguez
Bayesian Regression for Predicting Subscription to Bank Term Deposits in Direct Marketing Campaigns
Muhammad Farhan Tanvir, Md Maruf Hossain, Md Asifuzzaman Jishan
Accelerated Bayesian parameter estimation and model selection for gravitational waves with normalizing flows
Alicja Polanska, Thibeau Wouters, Peter T. H. Pang, Kaze K. W. Wong, Jason D. McEwen