Discovering better theories
Theories of human decision-making have proliferated in recent years. However, these theories are often difficult to distinguish from each other and offer limited improvement in accounting for patterns in decision-making over earlier theories. Peterson et al. leverage machine learning to evaluate classical decision theories, increase their predictive power, and generate new theories of decision-making (see the Perspective by Bhatia and He). This method has implications for theory generation in other domains.
Science, abe2629, this issue p. 1209; see also abi7668, p. 1150
Predicting and understanding how people make decisions has been a long-standing goal in many fields, with quantitative models of human decision-making informing research in both the social sciences and engineering. We show how progress toward this goal can be accelerated by using large datasets to power machine-learning algorithms that are constrained to produce interpretable psychological theories. Conducting the largest experiment on risky choice to date and analyzing the results using gradient-based optimization of differentiable decision theories implemented through artificial neural networks, we were able to recapitulate historical discoveries, establish that there is room to improve on existing theories, and discover a new, more accurate model of human decision-making in a form that preserves the insights from centuries of research.