The LCFI website uses cookies only for anonymised website statistics and for ensuring our security, never for tracking or identifying you individually. To find out more, and to find out how we protect your personal information, please read our privacy policy.

Artificial Intelligence and the Common Sense of Animals

Academic Journal article by Murray Shanahan, Matthew Crosby, Benjamin Beyret, Lucy Cheke

Artificial Intelligence and the Common Sense of Animals. Trends in Cognitive Sciences (TICS 2085) DOI:https://doi.org/10.1016/j.tics.2020.09.002

Abstract: The problem of common sense remains a major obstacle to progress in artificial intelligence. Here, we argue that common sense in humans is founded on a set of basic capacities that are possessed by many other animals, capacities pertaining to the understanding of objects, space, and causality. The field of animal cognition has developed numerous experimental protocols for studying these capacities and, thanks to progress in deep reinforcement learning (RL), it is now possible to apply these methods directly to evaluate RL agents in 3D environments. Besides evaluation, the animal cognition literature offers a rich source of behavioural data, which can serve as inspiration for RL tasks and curricula.

Download Academic Journal article