CNS*2020 Online has ended
Welcome to the Sched instance for CNS*2020 Online! Please read the instruction document on detailed information on CNS*2020.
Back To Schedule
Sunday, July 19 • 8:00pm - 9:00pm
P125: A predictive coding model of transitive inference

Log in to save this to your schedule, view media, leave feedback and see who's attending!

Feedback form is now closed.

Moritz Moeller
, Rafal Bogacz, Sanjay Manohar

Transitive inference--deducing that "A is better than C" from the premises "A is better than B" and "B is better than C"--is a basic form of deductive reasoning; both humans and animals are capable of it. However, the mechanism that enables transitive inference is not understood. Partly, this is due to the absence of a concrete, falsifiable formulation of the so-called cognitive explanation of transitive inference (which suggests that subjects combine the facts they observe into a mental model, which they then use for reasoning). In this work, we use the predictive coding method to derive a precise, mathematical implementation of the cognitive explanation of transitive inference (Fig. 1A shows a schematic representation of the model we use). We test our model by simulating a set of typical transitive inference experiments and show that it reproduces several phenomena observed in animal experiments. For example, our model reproduces the gradual acquisition of premise pairs (A > B, B > C) and the simultaneously emerging capability for transitive inference (A>C) (Fig. 1B). We expect this work to lead to novel testable predictions that will inspire future experiments and help to uncover the mechanism behind transitive inference. Further, our work adds support to predictive coding as a universal organising principle of brain function.

 If problems arise, email: moritz.moeller@stx.ox.ac.uk

avatar for Moritz Moeller

Moritz Moeller

PhD student, NDCN, University of Oxford

Sunday July 19, 2020 8:00pm - 9:00pm CEST
Slot 07