HackerPapers

arXiv:1611.01578cs.LG

Neural Architecture Search with Reinforcement Learning

Barret Zoph, Quoc V. Le

Abstract

Neural networks are powerful and flexible models that work well for many difficult learning tasks in image, speech and natural language understanding. Despite their success, neural networks are still hard to design. In this paper, we use a recurrent network to generate the model descriptions of neural networks and train this RNN with reinforcement learning to maximize the expected accuracy of the generated architectures on a validation set. On the CIFAR-10 dataset, our method, starting from scratch, can design a novel network architecture that rivals the best human-invented architecture in terms of test set accuracy. Our CIFAR-10 model achieves a test error rate of 3.65, which is 0.09 percent better and 1.05x faster than the previous state-of-the-art model that used a similar architectural scheme. On the Penn Treebank dataset, our model can compose a novel recurrent cell that outperforms the widely-used LSTM cell, and other state-of-the-art baselines. Our cell achieves a test set perplexity of 62.4 on the Penn Treebank, which is 3.6 perplexity better than the previous state-of-the-art model. The cell can also be transferred to the character language modeling task on PTB and achieves a state-of-the-art perplexity of 1.214.

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

On Hacker News

Linked by 10 identifiable commenters in 10 comments across 9 threads, from Dec 2016 to Jul 2018. Found in 1 HN submission from Jan 2017 to Jan 2017. Open submission discussions.

Comments linking this paper, per year.

Discussed in comments9 threads

  1. Darts: Differentiable architecture search for convolutional, recurrent networks 2 comments

    • Cool, was thinking about this subject earlier today. Abstract: >This paper addresses the scalability challenge of architecture search by formulating the task in a differentiable manner. Unlike conventional approaches of applying evolution or reinforcement learning over a…

      ashelmireView on HNParent context

    • Not GP, but you are probably looking for the "Neural Architecture Search" series [1] [2] [3]. First one uses something like 1k GPUs for a month, next one is a bit more reasonable, and the last one actually has a reasonable training time, and has comparable performance to DARTS…

      kripkeView on HNParent context

  2. Deep learning: a critical appraisal 1 comment

    • This question isn't silly! People have actually trained neural networks to generate other neural networks. The "designer" network is trained using reinforcement learning, with a reward signal that depends on the efficacy of the networks it creates. See: Neural Architecture…

      RoboTeddyView on HNParent context

  3. Using Google Cloud AutoML to Classify Poisonous Australian Spiders 1 comment

    • Jeff Dean talks about AutoML using RL and in the paper " Neural Architecture Search with Reinforcement Learning" it also talks about this. Also it seems different from more traditional hyperparameter optimization because it makes novel cells. So the structure of the network…

      michaelgreenView on HNParent context

  4. Building A.I. That Can Build A.I 1 comment

    • Yeah, I'm confused that this is the top comment; it's factually incorrect. NASNet is an example of a result of AutoML. To quote the Google blogpost on NASNet: >In Learning Transferable Architectures for Scalable Image Recognition, we apply AutoML to the ImageNet image…

      ollinView on HNParent context

  5. Hacker's guide to Neural Networks (2012) 1 comment

    • There has been some work related to what you are describing: Learning to learn by gradient descent by gradient descent https://arxiv.org/abs/1606.04474 Related (but less so), there are also some papers about learning neural network architectures: Designing Neural Network…

      chriskananView on HNParent context

  6. Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks 1 comment
  7. Ask HN: What is the most exciting development in your field right now? 1 comment

    • Deep learning architectures built by machines (so we no longer have to design architecture to solve problems) https://arxiv.org/abs/1611.01578 Transfer Learning (so we need less data to build models) http://ftp.cs.wisc.edu/machine-learning/shavlik-group/torrey... Generative…

      sarthakjainView on HNParent context

  8. Learning to Learn in TensorFlow 1 comment

    • Other equally exciting papers that relates to learning to learn in DL. "Neural Architecture Search with Reinforcement Learning" https://arxiv.org/abs/1611.01578 "RL^2: Fast Reinforcement Learning via Slow Reinforcement Learning" https://arxiv.org/abs/1611.02779 "Designing Neural…

      lucidrainsView on HNParent context

  9. This AI Boom Will Also Bust 1 comment

Submitted to Hacker News

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