Task-Relevant Feature Selection with Prediction Focused Mixture Models
Abstract
Probabilistic models, such as mixture models, can encode latent structures that both explain the data and aid specific downstream tasks. We focus on a constrained setting where we want to learn a model with relatively few components (e.g. for interpretability). Simultaneously, we ensure that the components are useful for downstream predictions by introducing \emph{prediction-focused} modeling for mixtures, which automatically selects data features relevant to a prediction task. Our approach identifies task-relevant input features, outperforms models that are not prediction-focused, and is easy to optimize; most importantly, we also characterize \emph{when} prediction-focused modeling can be expected to work.
Cite
Text
Sharma et al. "Task-Relevant Feature Selection with Prediction Focused Mixture Models." Transactions on Machine Learning Research, 2024.Markdown
[Sharma et al. "Task-Relevant Feature Selection with Prediction Focused Mixture Models." Transactions on Machine Learning Research, 2024.](https://mlanthology.org/tmlr/2024/sharma2024tmlr-taskrelevant/)BibTeX
@article{sharma2024tmlr-taskrelevant,
title = {{Task-Relevant Feature Selection with Prediction Focused Mixture Models}},
author = {Sharma, Abhishek and Zeng, Catherine and Narayanan, Sanjana and Parbhoo, Sonali and Perlis, Roy H. and Doshi-Velez, Finale},
journal = {Transactions on Machine Learning Research},
year = {2024},
url = {https://mlanthology.org/tmlr/2024/sharma2024tmlr-taskrelevant/}
}