Timezone: »
A central problem to understanding intelligence is the concept of generalisation. This allows previously learnt structure to be exploited to solve tasks in novel situations differing in their particularities. We take inspiration from neuroscience, specifically the hippocampal-entorhinal system known to be important for generalisation. We propose that to generalise structural knowledge, the representations of the structure of the world, i.e. how entities in the world relate to each other, need to be separated from representations of the entities themselves. We show, under these principles, artificial neural networks embedded with hierarchy and fast Hebbian memory, can learn the statistics of memories and generalise structural knowledge. Spatial neuronal representations mirroring those found in the brain emerge, suggesting spatial cognition is an instance of more general organising principles. We further unify many entorhinal cell types as basis functions for constructing transition graphs, and show these representations effectively utilise memories. We experimentally support model assumptions, showing a preserved relationship between entorhinal grid and hippocampal place cells across environments.
Author Information
James Whittington (University of Oxford)
Timothy Muller (University of Oxford)
Shirely Mark (University College London)
Caswell Barry (University College London)
Tim Behrens (University of Oxford)
Related Events (a corresponding poster, oral, or spotlight)
-
2018 Oral: Generalisation of structural knowledge in the hippocampal-entorhinal system »
Tue. Dec 4th 08:50 -- 09:05 PM Room Room 220 CD
More from the Same Authors
-
2019 Poster: Better Transfer Learning with Inferred Successor Maps »
Tamas Madarasz · Tim Behrens -
2019 Spotlight: Better Transfer Learning with Inferred Successor Maps »
Tamas Madarasz · Tim Behrens