Human-level concept learning through probabilistic program induction

@article{Lake2015HumanlevelCL,
  title={Human-level concept learning through probabilistic program induction},
  author={B. Lake and R. Salakhutdinov and J. Tenenbaum},
  journal={Science},
  year={2015},
  volume={350},
  pages={1332 - 1338}
}
  • B. Lake, R. Salakhutdinov, J. Tenenbaum
  • Published 2015
  • Computer Science, Medicine
  • Science
  • People learning new concepts can often generalize successfully from just a single example, yet machine learning algorithms typically require tens or hundreds of examples to perform with similar accuracy. [...] Key Method The model represents concepts as simple programs that best explain observed examples under a Bayesian criterion. On a challenging one-shot classification task, the model achieves human-level performance while outperforming recent deep learning approaches. We also present several “visual Turing…Expand Abstract
    1,395 Citations
    How intelligent are convolutional neural networks?
    • 5
    • PDF
    Bongard-LOGO: A New Benchmark for Human-Level Concept Learning and Reasoning
    • 1
    • Highly Influenced
    • PDF
    People Infer Recursive Visual Concepts from Just a Few Examples
    • 7
    • PDF
    Learning a Deep Generative Model like a Program: the Free Category Prior
    Fast Parametric Learning with Activation Memorization
    • 33
    • Highly Influenced
    • PDF
    Learning abstract visual concepts via probabilistic program induction in a Language of Thought
    • 14
    • Highly Influenced
    • PDF
    Theory-based learning in humans and machines
    The Omniglot Jr. challenge; Can a model achieve child-level character generation and classification?
    • 2020

    References

    SHOWING 1-10 OF 130 REFERENCES
    Using Generative Models for Handwritten Digit Recognition
    • 180
    • PDF
    Learning with Hierarchical-Deep Models
    • 190
    • PDF
    A Model of Inductive Bias Learning
    • J. Baxter
    • Computer Science
    • J. Artif. Intell. Res.
    • 2000
    • 740
    • PDF
    The Structure of Perceptual Categories
    • Feldman
    • Mathematics, Medicine
    • Journal of mathematical psychology
    • 1997
    • 81
    • PDF
    Gradient-based learning applied to document recognition
    • 25,611
    • PDF
    Dropout: a simple way to prevent neural networks from overfitting
    • 20,150
    • PDF
    Human-level control through deep reinforcement learning
    • 10,419
    • PDF
    A theory of the learnable
    • 3,566
    • PDF
    Object name Learning Provides On-the-Job Training for Attention
    • 480
    • PDF