Developing realistic models of human intelligence and learning is a major aspiration of several scholarly fields, including Artificial Intelligence, Economics, and Philosophy. The ability to conceptualize causal structure in the external world and to use causal representations to guide behavior is an essential feature of human intelligence. Indeed, causality is now the mainstream focus within a heterogeneous set of scientific circles. A review of prestigious research awards is one indicator of the breadth and depth of interest in the study of causal systems. For instance, in Computer Science, the 2011 Turing Award went to Judea Pearl for “the development of a calculus for causal reasoning.” In Economics, the 2021 Nobel Prize went to David Card, Guido Imbens, and Joshua Angrist for “advances in causal inference in economic settings.” In Machine Learning, there is a growing recognition that the path to building artificial agents that exhibit truly human-like intelligence runs through causal reasoning. For example, Yoshua Bengio, himself a recent winner of the Turing Award for his work in Deep Learning, says: "We don't have AI systems which actually understand at the level that humans do or anywhere close. What does understanding mean? ... In part it means capturing the causal structure in the world."

The goal of this initiative is to stimulate fruitful conversations about causal learning among a diverse community of world-class researchers. The recent burst of interdisciplinary research on causality and machine learning clearly shows the potential for mutual benefit through idea-sharing and integration. Here, we aim to bring two additional disciplines into the conversation: the first is game theory, which has much to say about interactive decision making, as well as having an extensive line of work exploring a broad range of human learning dynamics. The second is philosophy, which has much to say about human conceptualization, certain lines of which are particularly relevant to interactive causal learning. We are casting a wide net to draw in a broad range of cutting-edge perspectives - be it work investigating specific aspects of causality per se, or cutting-edge work that may not be directly about causality but, nevertheless, holds promise to benefit our effort.

CONFERENCE ORGANIZERS