Year: 2023
Type: article
Abstract: In real applications, interaction between machine learning models and domain experts is critical; however, the classical machine learning paradigm that usually produces only a single model does not facilitate such interaction. Approximating and exploring the Rashomon set, i.e., the set of all near-optimal models, addresses this practical challenge by providing the user with a searchable space containing a diverse set of models from which domain experts can choose. We present algorithms to efficiently and accurately approximate the Rashomon set of sparse, generalized additive models with ellipsoids for fixed support sets and use these ellipsoids to approximate Rashomon sets for many different support sets. The approximated Rashomon set serves as a cornerstone to solve practical challenges such as (1) studying the variable importance for the model class; (2) finding models under user-specified constraints (monotonicity, direct editing); and (3) investigating sudden changes in the shape functions. Experiments demonstrate the fidelity of the approximated Rashomon set and its effectiveness in solving practical challenges.
Source: PubMed
Institutions: Duke University, University of British Columbia
Language: English
Subfield: Statistics and Probability
Field: Mathematics
Domain: Physical Sciences
Open Access status: green
Awards: FAIR DATA AND INTERPRETABLE AI FRAMEWORK FOR ARCHITECTURED METAMATERIALS, EXPLORING THE WHOLE SET OF SPARSE EXPLANATIONS, G6848452693, NRT-HDR: Harnessing AI for Understanding & Designing Materials (aiM), EAGER: Creating an Unsupervised Interpretable Representation of the World Through Concept Disentanglement
OpenAlex ID: W4361229290