is our passion
Home / Projects / The Computer Science of Human Decisions

The Computer Science of Human Decisions

NOMIS Project 2021

— 2026

How do people trade off risk and reward? How do people decide how long they are willing to wait for a good thing? How do societies converge on good strategies for making decisions? For a long time, answering these questions has been the project of psychologists, economists and sociologists. However, the answers to these questions are becoming increasingly important to other scientists. An engineer trying to build a self-driving car needs to understand the decisions that people make on the road. A doctor trying to predict the course of a pandemic needs to know when people will make a short-term sacrifice for a long-term gain. Both of them need high-precision models of human decision-making.

Even as the study of decision-making begins to have an impact on a wider range of scientific disciplines, those disciplines are beginning to offer new tools for making sense of human behavior. In particular, recent advances in computer science have resulted in software that makes it possible to collect data at unprecedented scales and machine learning methods that can be used to automatically identify the patterns in those data. These technologies create a unique opportunity to study the human mind in a new way.

The Computer Science of Human Decisions project aims to seize that opportunity, integrating computer science with psychology to develop high-precision models of decision-making. The outcome of this research project will not just be better models for predicting human decisions, but a deeper integration of the classic tools of the social and behavioral sciences with those of computer science.

The project is being led by Tom Griffiths at Princeton University (Princeton, US).


NOMIS Researcher(s)

Henry R. Luce Professor of Information Technology, Consciousness and Culture
Princeton University

Project News

The insights of NOMIS researcher Tom Griffiths and colleagues have been featured in an article in The Guardian that explores the changing approach to knowledge acquisition. Griffiths is leading The […]

NOMIS researcher Tom Griffiths and colleagues have shown how they can leverage machine learning to evaluate classical decision theories, increase their predictive power, and generate new theories of decision-making. Their […]


Project Insights


Effectively updating one’s beliefs requires sufficient empirical evidence (i.e., data) and the computational capacity to process it. Yet both data and computational resources are limited for human minds. Here, we […]

Abstract: Large-scale social networks are thought to contribute to polarization by amplifying people’s biases. However, the complexity of these technologies makes it difficult to identify the mechanisms responsible and evaluate mitigation strategies. Here we show under controlled laboratory conditions that transmission through social networks amplifies motivational biases on a simple artificial
Abstract: Supervised learning typically focuses on learning transferable representations from training examples annotated by humans. While rich annotations (like soft labels) carry more information than sparse annotations (like hard labels), they are also more expensive to collect. For example, while hard labels only provide information about the closest class an object