Matthew Botvinick

TL;DR: In this paper, the authors used fMRI to measure predictive representations in a setting where the Successor Representation implies specific errors in multi-step expectancies and corresponding behavioral errors.

TL;DR: A novel architecture, MEMO, endowed with the capacity to reason over longer distances is developed with the addition of two novel components, which introduces a separation between memories/facts stored in external memory and the items that comprise these facts in External memory.

TL;DR: A recursive implicit planning module is developed that operates over episodic memories, and it is shown that the resulting deep-RL agent is able to explore and plan in novel environments, outperforming the nearest baseline by factors of 2-3 across the two domains.

TL;DR: The authors reply here, addressing both Bowers et al.'s criticisms of the Botvinick and Plaut model and the former's assessment of parallel distributed processing models in general.

TL;DR: A new method is presented for assessing the degree to which CDSM capture semantic interactions that dissociates the influences of lexical and compositional information and shows that neural language input vectors are consistently superior to co-occurrence based vectors and vector addition matches and is in many cases superior to purpose-built paramaterized models.