The First International Workshop on Computational Models of the Visual Cortex: Hierarchies, Layers, Sparsity, Saliency and Attention

Research Article

Learning Abstract Classes using Deep Learning

  • @INPROCEEDINGS{10.4108/eai.3-12-2015.2262468,
        author={Sebastian Stabinger and Antonio Rodr\^{\i}guez-S\^{a}nchez and Justus Piater},
        title={Learning Abstract Classes using Deep Learning},
        proceedings={The First International Workshop on Computational Models of the Visual Cortex: Hierarchies, Layers, Sparsity, Saliency and Attention},
        publisher={ACM},
        proceedings_a={CMVC},
        year={2016},
        month={5},
        keywords={deep learning convolutional neural networks visual cortex abstract reasoning},
        doi={10.4108/eai.3-12-2015.2262468}
    }
    
  • Sebastian Stabinger
    Antonio Rodríguez-Sánchez
    Justus Piater
    Year: 2016
    Learning Abstract Classes using Deep Learning
    CMVC
    ACM
    DOI: 10.4108/eai.3-12-2015.2262468
Sebastian Stabinger1,*, Antonio Rodríguez-Sánchez1, Justus Piater1
  • 1: University of Innsbruck
*Contact email: Sebastian.Stabinger@uibk.ac.at

Abstract

Humans are generally good at learning abstract concepts about objects and scenes (e.g. spatial orientation, relative sizes, etc.). Over the last years convolutional neural networks have achieved almost human performance in recognizing concrete classes (i.e. specific object categories). This paper tests the performance of a current CNN (GoogLeNet) on the task of differentiating between abstract classes which are trivially differentiable for humans. We trained and tested the CNN on the two abstract classes of horizontal and vertical orientation and determined how well the network is able to transfer the learned classes to other, previously unseen objects.