DOE PAGES title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Leveraging Prior Concept Learning Improves Generalization From Few Examples in Computational Models of Human Object Recognition

Abstract

Humans quickly and accurately learn new visual concepts from sparse data, sometimes just a single example. The impressive performance of artificial neural networks which hierarchically pool afferents across scales and positions suggests that the hierarchical organization of the human visual system is critical to its accuracy. These approaches, however, require magnitudes of order more examples than human learners. We used a benchmark deep learning model to show that the hierarchy can also be leveraged to vastly improve the speed of learning. We specifically show how previously learned but broadly tuned conceptual representations can be used to learn visual concepts from as few as two positive examples; reusing visual representations from earlier in the visual hierarchy, as in prior approaches, requires significantly more examples to perform comparably. These results suggest techniques for learning even more efficiently and provide a biologically plausible way to learn new visual concepts from few examples.

Authors:
;
Publication Date:
Sponsoring Org.:
USDOE
OSTI Identifier:
1759144
Grant/Contract Number:  
COMP-19- ERD-007; AC52-07NA27344
Resource Type:
Published Article
Journal Name:
Frontiers in Computational Neuroscience
Additional Journal Information:
Journal Name: Frontiers in Computational Neuroscience Journal Volume: 14; Journal ID: ISSN 1662-5188
Publisher:
Frontiers Media SA
Country of Publication:
Switzerland
Language:
English

Citation Formats

Rule, Joshua S., and Riesenhuber, Maximilian. Leveraging Prior Concept Learning Improves Generalization From Few Examples in Computational Models of Human Object Recognition. Switzerland: N. p., 2021. Web. doi:10.3389/fncom.2020.586671.
Rule, Joshua S., & Riesenhuber, Maximilian. Leveraging Prior Concept Learning Improves Generalization From Few Examples in Computational Models of Human Object Recognition. Switzerland. https://doi.org/10.3389/fncom.2020.586671
Rule, Joshua S., and Riesenhuber, Maximilian. Tue . "Leveraging Prior Concept Learning Improves Generalization From Few Examples in Computational Models of Human Object Recognition". Switzerland. https://doi.org/10.3389/fncom.2020.586671.
@article{osti_1759144,
title = {Leveraging Prior Concept Learning Improves Generalization From Few Examples in Computational Models of Human Object Recognition},
author = {Rule, Joshua S. and Riesenhuber, Maximilian},
abstractNote = {Humans quickly and accurately learn new visual concepts from sparse data, sometimes just a single example. The impressive performance of artificial neural networks which hierarchically pool afferents across scales and positions suggests that the hierarchical organization of the human visual system is critical to its accuracy. These approaches, however, require magnitudes of order more examples than human learners. We used a benchmark deep learning model to show that the hierarchy can also be leveraged to vastly improve the speed of learning. We specifically show how previously learned but broadly tuned conceptual representations can be used to learn visual concepts from as few as two positive examples; reusing visual representations from earlier in the visual hierarchy, as in prior approaches, requires significantly more examples to perform comparably. These results suggest techniques for learning even more efficiently and provide a biologically plausible way to learn new visual concepts from few examples.},
doi = {10.3389/fncom.2020.586671},
journal = {Frontiers in Computational Neuroscience},
number = ,
volume = 14,
place = {Switzerland},
year = {Tue Jan 12 00:00:00 EST 2021},
month = {Tue Jan 12 00:00:00 EST 2021}
}

Works referenced in this record:

Robust Object Recognition with Cortex-Like Mechanisms
journal, March 2007

  • Serre, Thomas; Wolf, Lior; Bileschi, Stanley
  • IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 29, Issue 3
  • DOI: 10.1109/TPAMI.2007.56

What the left and right anterior fusiform gyri tell us about semantic memory
journal, October 2010


What are the visual features underlying rapid object recognition?
journal, January 2011


Performance-optimized hierarchical models predict neural responses in higher visual cortex
journal, May 2014

  • Yamins, D. L. K.; Hong, H.; Cadieu, C. F.
  • Proceedings of the National Academy of Sciences, Vol. 111, Issue 23
  • DOI: 10.1073/pnas.1403112111

The neural and computational bases of semantic cognition
journal, November 2016

  • Ralph, Matthew A. Lambon; Jefferies, Elizabeth; Patterson, Karalyn
  • Nature Reviews Neuroscience, Vol. 18, Issue 1
  • DOI: 10.1038/nrn.2016.150

The role of conceptual knowledge in object use Evidence from semantic dementia
journal, September 2000


The Neurobiology of Category Learning
journal, June 2004

  • Ashby, F. Gregory; Spiering, Brian J.
  • Behavioral and Cognitive Neuroscience Reviews, Vol. 3, Issue 2
  • DOI: 10.1177/1534582304270782

Learning multisensory representations for auditory-visual transfer of sequence category knowledge: a probabilistic language of thought approach
journal, October 2014


First-Pass Selectivity for Semantic Categories in Human Anteroventral Temporal Lobe
journal, December 2011


ImageNet Large Scale Visual Recognition Challenge
journal, April 2015

  • Russakovsky, Olga; Deng, Jia; Su, Hao
  • International Journal of Computer Vision, Vol. 115, Issue 3
  • DOI: 10.1007/s11263-015-0816-y

Human-level concept learning through probabilistic program induction
journal, December 2015


Hierarchical models of object recognition in cortex
journal, November 1999

  • Riesenhuber, Maximilian; Poggio, Tomaso
  • Nature Neuroscience, Vol. 2, Issue 11
  • DOI: 10.1038/14819

Models of object recognition
journal, November 2000

  • Riesenhuber, Maximilian; Poggio, Tomaso
  • Nature Neuroscience, Vol. 3, Issue S11
  • DOI: 10.1038/81479

Building machines that learn and think like people
journal, November 2016

  • Lake, Brenden M.; Ullman, Tomer D.; Tenenbaum, Joshua B.
  • Behavioral and Brain Sciences, Vol. 40
  • DOI: 10.1017/S0140525X16001837

Where Is the Semantic System? A Critical Review and Meta-Analysis of 120 Functional Neuroimaging Studies
journal, March 2009

  • Binder, Jeffrey R.; Desai, Rutvik H.; Graves, William W.
  • Cerebral Cortex, Vol. 19, Issue 12
  • DOI: 10.1093/cercor/bhp055

ImageNet: A large-scale hierarchical image database
conference, June 2009

  • Deng, Jia; Dong, Wei; Socher, Richard
  • 2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops (CVPR Workshops), 2009 IEEE Conference on Computer Vision and Pattern Recognition
  • DOI: 10.1109/CVPR.2009.5206848

Transfer of object category knowledge across visual and haptic modalities: Experimental and computational studies
journal, February 2013


Evaluation of a Shape-Based Model of Human Face Discrimination Using fMRI and Behavioral Techniques
journal, April 2006


CNN Features Off-the-Shelf: An Astounding Baseline for Recognition
conference, June 2014

  • Razavian, Ali Sharif; Azizpour, Hossein; Sullivan, Josephine
  • 2014 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)
  • DOI: 10.1109/CVPRW.2014.131

Multivariate Pattern Analysis Reveals Category-Related Organization of Semantic Representations in Anterior Temporal Cortex
journal, September 2016


Time Course of Shape and Category Selectivity Revealed by EEG Rapid Adaptation
journal, February 2014

  • Scholl, Clara A.; Jiang, Xiong; Martin, Jacob G.
  • Journal of Cognitive Neuroscience, Vol. 26, Issue 2
  • DOI: 10.1162/jocn_a_00477

The ventral visual pathway: an expanded neural framework for the processing of object quality
journal, January 2013

  • Kravitz, Dwight J.; Saleem, Kadharbatcha S.; Baker, Chris I.
  • Trends in Cognitive Sciences, Vol. 17, Issue 1
  • DOI: 10.1016/j.tics.2012.10.011

A review of visual memory capacity: Beyond individual items and toward structured representations
journal, May 2011

  • Brady, T. F.; Konkle, T.; Alvarez, G. A.
  • Journal of Vision, Vol. 11, Issue 5
  • DOI: 10.1167/11.5.4

Creating Concepts from Converging Features in Human Cortex
journal, March 2014

  • Coutanche, Marc N.; Thompson-Schill, Sharon L.
  • Cerebral Cortex, Vol. 25, Issue 9
  • DOI: 10.1093/cercor/bhu057

A feedforward architecture accounts for rapid categorization
journal, April 2007

  • Serre, T.; Oliva, A.; Poggio, T.
  • Proceedings of the National Academy of Sciences, Vol. 104, Issue 15, p. 6424-6429
  • DOI: 10.1073/pnas.0700622104

Investigating Causal Relations by Econometric Models and Cross-spectral Methods
journal, August 1969


The reverse hierarchy theory of visual perceptual learning
journal, October 2004


Visual long-term memory has a massive storage capacity for object details
journal, September 2008

  • Brady, T. F.; Konkle, T.; Alvarez, G. A.
  • Proceedings of the National Academy of Sciences, Vol. 105, Issue 38
  • DOI: 10.1073/pnas.0803390105

A Comparison of Primate Prefrontal and Inferior Temporal Cortices during Visual Categorization
journal, June 2003


Attention, similarity, and the identification–categorization relationship.
journal, January 1986


Rapid consolidation of new knowledge in adulthood via fast mapping
journal, September 2015


The computational magic of the ventral stream
journal, March 2012


Fast mapping rapidly integrates information into existing memory networks.
journal, January 2014

  • Coutanche, Marc N.; Thompson-Schill, Sharon L.
  • Journal of Experimental Psychology: General, Vol. 143, Issue 6
  • DOI: 10.1037/xge0000020

Learning abstract visual concepts via probabilistic program induction in a Language of Thought
journal, November 2017


Categorization Training Results in Shape- and Category-Selective Human Neural Plasticity
journal, March 2007


The neurobiology of semantic memory
journal, November 2011


The representational dynamics of task and object processing in humans
journal, January 2018


Learning and Transferring Mid-level Image Representations Using Convolutional Neural Networks
conference, June 2014

  • Oquab, Maxime; Bottou, Leon; Laptev, Ivan
  • 2014 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
  • DOI: 10.1109/CVPR.2014.222

Functional anatomy of a common semantic system for words and pictures
journal, September 1996

  • Vandenberghe, R.; Price, C.; Wise, R.
  • Nature, Vol. 383, Issue 6597
  • DOI: 10.1038/383254a0

Rapid neocortical acquisition of long-term arbitrary associations independent of the hippocampus
journal, January 2011

  • Sharon, T.; Moscovitch, M.; Gilboa, A.
  • Proceedings of the National Academy of Sciences, Vol. 108, Issue 3
  • DOI: 10.1073/pnas.1005238108

Granger Causality Analysis in Neuroscience and Neuroimaging
journal, February 2015


Connectivity-based constraints on category-specificity in the ventral object processing pathway
journal, October 2017


Encoding of Categories by Noncategory-Specific Neurons in the Inferior Temporal Cortex
journal, February 2001

  • Thomas, Elizabeth; Van Hulle, Marc M.; Vogel, Rufin
  • Journal of Cognitive Neuroscience, Vol. 13, Issue 2
  • DOI: 10.1162/089892901564252