HackerPapers

arXiv:1410.4615cs.NE

Learning to Execute

Wojciech Zaremba, Ilya Sutskever

Abstract

Recurrent Neural Networks (RNNs) with Long Short-Term Memory units (LSTM) are widely used because they are expressive and are easy to train. Our interest lies in empirically evaluating the expressiveness and the learnability of LSTMs in the sequence-to-sequence regime by training them to evaluate short computer programs, a domain that has traditionally been seen as too complex for neural networks. We consider a simple class of programs that can be evaluated with a single left-to-right pass using constant memory. Our main result is that LSTMs can learn to map the character-level representations of such programs to their correct outputs. Notably, it was necessary to use curriculum learning, and while conventional curriculum learning proved ineffective, we developed a new variant of curriculum learning that improved our networks' performance in all experimental conditions. The improved curriculum had a dramatic impact on an addition problem, making it possible to train an LSTM to add two 9-digit numbers with 99% accuracy.

arXiv comments
Journal reference
DOI
Subjects
cs.NENeural and Evolutionary Computing cs.AIArtificial Intelligence cs.LGMachine Learning

On Hacker News

Linked by 7 identifiable commenters in 7 comments across 7 threads, from Oct 2014 to Jun 2017. Found in 2 HN submissions from Oct 2014 to Nov 2014. Open submission discussions.

Comments linking this paper, per year.

Discussed in comments7 threads

  1. DeepFix: Fixing Common C Language Errors by Deep Learning 1 comment

    • There was similar work by W. Zaremba (OpenAI I think). LSTM trained to interpret python source code. Took a ton of hackery to get it to not fall apart on simple operations and it continued making indexing errors. I wouldn't read too much into these results.…

      shriphaniView on HNParent context

  2. The future is fewer people writing code? 1 comment

    • There are signs, if you know were to look for them: http://arxiv.org/abs/1510.07211 https://arxiv.org/abs/1410.4615 http://arxiv.org/abs/1605.06640 https://arxiv.org/abs/1410.5401

      sapphireblueView on HNParent context

  3. A Neural Conversational Model [pdf] 1 comment

    • Whilst you are correct that the number of allowed computation steps is limited, LSTMs have still been used with success for such tasks. LSTMs have been used to perform addition with 99% accuracy^ on two 9-digit numbers[1]. The paper even shows that simple Python programs can be…

      SmerityView on HNParent context

  4. The Unreasonable Effectiveness of Recurrent Neural Networks 1 comment

    • In "Learning to Execute" by Zaremba & Sutskever http://arxiv.org/abs/1410.4615 An RNN learns snippets of python Their next paper is "Reinforcement Learning Neural Turing Machines" http://arxiv.org/abs/1505.00521 based on Graves "Neural Turing Machines"…

      deepnetView on HNParent context

  5. Learning to Execute [pdf] 1 comment

    • No big criticism implied, but some of us like to see the abstract, first. So here's the link to it: http://arxiv.org/abs/1410.4615

      pasbesoinView on HNParent context

  6. Jeff Hawkins on the Limitations of Artificial Neural Networks 1 comment

    • > has built some impressive predictive systems Do you have any links to examples? I'm not aware of any systems built with HTMs that I'd consider "impressive" relative to what ANNs are doing these days, e.g. learning to play Atari games by watching the screen [1], approaching…

      modelessView on HNParent context

  7. Learning to Execute and Neural Turing Machines 1 comment

Submitted to Hacker News

Connected through shared HN threads, not inferred similarity or a recommendation.