Building machines that learn and think for themselves
Matthew Botvinick, David G. T. Barrett, Peter Battaglia, Nando de Freitas, Darshan Kumaran, Joel Z. Leibo, Timothy Lillicrap, Joseph Modayil, Shakir Mohamed, Neil C. Rabinowitz, Danilo Jimenez Rezende, Adam Santoro, Tom Schaul, Christopher Summerfield, Greg Wayne, Théophane Weber, Daan Wierstra, Shane Legg, Demis Hassabis · Behavioral and Brain Sciences · 2017
We agree with Lake and colleagues on their list of "key ingredients" for building human-like intelligence, including the idea that model-based reasoning is essential. However, we favor an approach that centers on one additional ingredient: autonomy. In particular, we aim toward agents that can both build and exploit their own internal models, with minimal human hand engineering. We believe an approach centered on autonomous learning has the greatest chance of success as we scale toward real-world complexity, tackling domains for which ready-made formal models are not available. Here, we survey several important examples of the progress that has been made toward building autonomous agents with human-like abilities, and highlight some outstanding challenges.