Model learning with personalized interpretability estimation (ML-PIE)
Paper i proceeding, 2021

High-stakes applications require AI-generated models to be interpretable. Current algorithms for the synthesis of potentially interpretable models rely on objectives or regularization terms that represent interpretability only coarsely (e.g., model size) and are not designed for a specific user. Yet, interpretability is intrinsically subjective. In this paper, we propose an approach for the synthesis of models that are tailored to the user by enabling the user to steer the model synthesis process according to her or his preferences. We use a bi-objective evolutionary algorithm to synthesize models with trade-offs between accuracy and a user-specific notion of interpretability. The latter is estimated by a neural network that is trained concurrently to the evolution using the feedback of the user, which is collected using uncertainty-based active learning. To maximize usability, the user is only asked to tell, given two models at the time, which one is less complex. With experiments on two real-world datasets involving 61 participants, we find that our approach is capable of learning estimations of interpretability that can be very different for different users. Moreover, the users tend to prefer models found using the proposed approach over models found using non-personalized interpretability indices.

explainable artificial intelligence

genetic programming

interpretable machine learning

active learning

neural networks

Författare

Marco Virgolin

Chalmers, Mekanik och maritima vetenskaper, Fordonsteknik och autonoma system

Andrea De Lorenzo

Universita degli Studi di Trieste

Francesca Randone

IMT Alti Studi Lucca

Eric Medvet

Universita degli Studi di Trieste

Mattias Wahde

Chalmers, Mekanik och maritima vetenskaper, Fordonsteknik och autonoma system

GECCO 2021 Companion - Proceedings of the 2021 Genetic and Evolutionary Computation Conference Companion

1355-1364

2021 Genetic and Evolutionary Computation Conference, GECCO 2021
Virtual, Online, France,

Ämneskategorier

Annan data- och informationsvetenskap

Bioinformatik (beräkningsbiologi)

Människa-datorinteraktion (interaktionsdesign)

DOI

10.1145/3449726.3463166

Mer information

Senast uppdaterat

2021-08-09