Abstract Reasoning with Distracting Features

Abstraction reasoning is a long-standing challenge in artificial intelligence. Recent studies suggest that many of the deep architectures that have triumphed over other domains failed to work well in abstract reasoning. In this paper, we first illustrate that one of the main challenges in such a reasoning task is the presence of distracting features, which requires the learning algorithm to leverage counter-evidence and to reject any of false hypothesis in order to learn the true patterns. We later show that carefully designed learning trajectory over different categories of training data can effectively boost learning performance by mitigating the impacts of distracting features. Inspired this fact, we propose feature robust abstract reasoning (FRAR) model, which consists of a reinforcement learning based teacher network to determine the sequence of training and a student network for predictions. Experimental results demonstrated strong improvements over baseline algorithms and we are able to beat the state-of-the-art models by 18.7\% in RAVEN dataset and 13.3\% in the PGM dataset.

[1]  Demis Hassabis,et al.  SCAN: Learning Abstract Hierarchical Compositional Visual Concepts , 2017, ArXiv.

[2]  Stephan Lewandowsky,et al.  A Bayesian Model of Rule Induction in Raven's Progressive Matrices , 2012, CogSci.

[3]  Shiguang Shan,et al.  Self-Paced Learning with Diversity , 2014, NIPS.

[4]  Ross B. Girshick,et al.  Focal Loss for Dense Object Detection , 2017, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[5]  Bernhard Schölkopf,et al.  Challenging Common Assumptions in the Unsupervised Learning of Disentangled Representations , 2018, ICML.

[6]  Bin Yang,et al.  Learning to Reweight Examples for Robust Deep Learning , 2018, ICML.

[7]  Razvan Pascanu,et al.  A simple neural network module for relational reasoning , 2017, NIPS.

[8]  Lawrence D. Jackel,et al.  Backpropagation Applied to Handwritten Zip Code Recognition , 1989, Neural Computation.

[9]  Christopher K. I. Williams,et al.  A Framework for the Quantitative Evaluation of Disentangled Representations , 2018, ICLR.

[10]  Alexei A. Efros,et al.  Ensemble of exemplar-SVMs for object detection and beyond , 2011, 2011 International Conference on Computer Vision.

[11]  Sjoerd van Steenkiste,et al.  Are Disentangled Representations Helpful for Abstract Visual Reasoning? , 2019, NeurIPS.

[12]  Tim Verbelen,et al.  Improving Generalization for Abstract Reasoning Tasks Using Disentangled Feature Representations , 2018, NIPS 2018.

[13]  Daphne Koller,et al.  Self-Paced Learning for Latent Variable Models , 2010, NIPS.

[14]  J. Stenton,et al.  Learning how to teach. , 1973, Nursing mirror and midwives journal.

[15]  Selmer Bringsjord,et al.  What is Artificial Intelligence? Psychometric AI as an Answer , 2003, IJCAI.

[16]  Felix Hill,et al.  Measuring abstract reasoning in neural networks , 2018, ICML.

[17]  Ran He,et al.  Self-Paced Learning: An Implicit Regularization Perspective , 2016, AAAI.

[18]  Kenneth D. Forbus,et al.  Modeling Visual Problem Solving as Analogical Reasoning , 2017, Psychological review.

[19]  Tom Eccles,et al.  Life-Long Disentangled Representation Learning with Cross-Domain Latent Homologies , 2018, NeurIPS.

[20]  J. Raven,et al.  Manual for Raven's progressive matrices and vocabulary scales , 1962 .

[21]  Felix Hill,et al.  Learning to Make Analogies by Contrasting Abstract Relational Structure , 2019, ICLR.

[22]  M A Just,et al.  From the SelectedWorks of Marcel Adam Just 1990 What one intelligence test measures : A theoretical account of the processing in the Raven Progressive Matrices Test , 2016 .

[23]  Kaiming He,et al.  Focal Loss for Dense Object Detection , 2017, 2017 IEEE International Conference on Computer Vision (ICCV).

[24]  Michael I. Jordan,et al.  Information Constraints on Auto-Encoding Variational Bayes , 2018, NeurIPS.

[25]  Andrew McCallum,et al.  Active Bias: Training More Accurate Neural Networks by Emphasizing High Variance Samples , 2017, NIPS.

[26]  Kenneth D. Forbus,et al.  A Structure-Mapping Model of Raven's Progressive Matrices , 2010 .

[27]  Hao Wu,et al.  Hierarchical Disentangled Representations , 2018, ArXiv.

[28]  Li Fei-Fei,et al.  MentorNet: Learning Data-Driven Curriculum for Very Deep Neural Networks on Corrupted Labels , 2017, ICML.

[29]  José Hernández-Orallo,et al.  Computer models solving intelligence test problems: Progress and implications , 2016, Artif. Intell..

[30]  Shiguang Shan,et al.  Self-Paced Curriculum Learning , 2015, AAAI.

[31]  Ron Sun,et al.  Similarity-Based Reasoning, Raven's Matrices, and General Intelligence , 2018, IJCAI.

[32]  Jason Weston,et al.  Curriculum learning , 2009, ICML '09.

[33]  Feng Gao,et al.  RAVEN: A Dataset for Relational and Analogical Visual REasoNing , 2019, 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[34]  Kenneth D. Forbus,et al.  Solving Geometric Analogy Problems Through Two-Stage Analogical Mapping , 2009, Cogn. Sci..

[35]  Abhishek Kumar,et al.  Variational Inference of Disentangled Latent Concepts from Unlabeled Observations , 2017, ICLR.

[36]  Jian Sun,et al.  Deep Residual Learning for Image Recognition , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[37]  Ashok K. Goel,et al.  Confident Reasoning on Raven's Progressive Matrices Tests , 2014, AAAI.

[38]  Michael Werman,et al.  IQ of Neural Networks , 2017, ArXiv.

[39]  Ke Wang,et al.  Automatic Generation of Raven's Progressive Matrices , 2015, IJCAI.

[40]  Klaus Greff,et al.  Multi-Object Representation Learning with Iterative Variational Inference , 2019, ICML.