A Self-Taught Artificial Agent for Multi-Physics Computational Model Personalization

TL;DR

This paper introduces Vito, a self-taught AI agent that uses reinforcement learning to automate and improve the personalization of multi-physics computational models in clinical settings, making the process more robust and efficient.

Contribution

The paper presents a model-independent reinforcement learning approach for model personalization, reducing the need for manual tuning and prior knowledge of the model.

Findings

Vito learned to optimize cost functions in synthetic scenarios.

Vito achieved comparable or better fit than standard methods in cardiac electrophysiology.

Vito demonstrated up to 11% higher success rates and 7 times faster convergence.

Abstract

Personalization is the process of fitting a model to patient data, a critical step towards application of multi-physics computational models in clinical practice. Designing robust personalization algorithms is often a tedious, time-consuming, model- and data-specific process. We propose to use artificial intelligence concepts to learn this task, inspired by how human experts manually perform it. The problem is reformulated in terms of reinforcement learning. In an off-line phase, Vito, our self-taught artificial agent, learns a representative decision process model through exploration of the computational model: it learns how the model behaves under change of parameters. The agent then automatically learns an optimal strategy for on-line personalization. The algorithm is model-independent; applying it to a new model requires only adjusting few hyper-parameters of the agent and defining the observations to match. The full knowledge of the model itself is not required. Vito was tested in a synthetic scenario, showing that it could learn how to optimize cost functions generically. Then Vito was applied to the inverse problem of cardiac electrophysiology and the personalization of a whole-body circulation model. The obtained results suggested that Vito could achieve equivalent, if not better goodness of fit than standard methods, while being more robust (up to 11% higher success rates) and with faster (up to seven times) convergence rate. Our artificial intelligence approach could thus make personalization algorithms generalizable and self-adaptable to any patient and any model.