A dataset and architecture for visual reasoning with a working memory

Guangyu Robert Yang, Igor Ganichev, Xiao Jing Wang, Jonathon Shlens, David Sussillo

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

A vexing problem in artificial intelligence is reasoning about events that occur in complex, changing visual stimuli such as in video analysis or game play. Inspired by a rich tradition of visual reasoning and memory in cognitive psychology and neuroscience, we developed an artificial, configurable visual question and answer dataset (COG) to parallel experiments in humans and animals. COG is much simpler than the general problem of video analysis, yet it addresses many of the problems relating to visual and logical reasoning and memory – problems that remain challenging for modern deep learning architectures. We additionally propose a deep learning architecture that performs competitively on other diagnostic VQA datasets (i.e. CLEVR) as well as easy settings of the COG dataset. However, several settings of COG result in datasets that are progressively more challenging to learn. After training, the network can zero-shot generalize to many new tasks. Preliminary analyses of the network architectures trained on COG demonstrate that the network accomplishes the task in a manner interpretable to humans.

Original languageEnglish (US)
Title of host publicationComputer Vision – ECCV 2018 - 15th European Conference, 2018, Proceedings
EditorsMartial Hebert, Vittorio Ferrari, Cristian Sminchisescu, Yair Weiss
PublisherSpringer Verlag
Pages729-745
Number of pages17
ISBN (Print)9783030012489
DOIs
StatePublished - Jan 1 2018
Event15th European Conference on Computer Vision, ECCV 2018 - Munich, Germany
Duration: Sep 8 2018Sep 14 2018

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11214 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other15th European Conference on Computer Vision, ECCV 2018
CountryGermany
CityMunich
Period9/8/189/14/18

    Fingerprint

Keywords

  • Recurrent network
  • Visual question answering
  • Visual reasoning
  • Working memory

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Yang, G. R., Ganichev, I., Wang, X. J., Shlens, J., & Sussillo, D. (2018). A dataset and architecture for visual reasoning with a working memory. In M. Hebert, V. Ferrari, C. Sminchisescu, & Y. Weiss (Eds.), Computer Vision – ECCV 2018 - 15th European Conference, 2018, Proceedings (pp. 729-745). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11214 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-030-01249-6_44