There is an error in equation 15, the “discounted time derivative” ht is defined incorrectly. It should read as follows:
“Operationally, the only change from the Kalman filter model described above is to replace the stimulus features xt with their discounted time derivative, ht = xt - γxt+1. To see why this makes sense, note that the immediate reward can be expressed in terms of the difference between two values:
(1)
This error does not affect the simulations, which were implemented with the correct definition.
Reference
- 1. Gershman SJ (2015) A Unifying Probabilistic View of Associative Learning. PLoS Comput Biol 11(11): e1004567. https://doi.org/10.1371/journal.pcbi.1004567 pmid:26535896
Citation: The PLOS Computational Biology Staff (2017) Correction: A Unifying Probabilistic View of Associative Learning. PLoS Comput Biol 13(11): e1005829. https://doi.org/10.1371/journal.pcbi.1005829
Published: November 16, 2017
Copyright: © 2017 The PLOS Computational Biology Staff. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.