Generative Teaching Networks: Accelerating Neural Architecture Search by Learning to Generate Synthetic Training Data

Abstract

This paper investigates the intriguing question of whether we can create learning algorithms that automatically generate training data, learning environments, and curricula in order to help AI agents rapidly learn. We show that such algorithms are possible via Generative Teaching Networks (GTNs), a general approach that is, in theory, applicable to supervised, unsupervised, and reinforcement learning, although our experiments only focus on the supervised case. GTNs are deep neural networks that generate data and/or training environments that a learner (e.g. a freshly initialized neural network) trains on for a few SGD steps before being tested on a target task. We then differentiate through the entire learning process via meta-gradients to update the GTN parameters to improve performance on the target task. GTNs have the beneficial property that they can theoretically generate any type of data or training environment, making their potential impact large. This paper introduces GTNs, discusses their potential, and showcases that they can substantially accelerate learning. We also demonstrate a practical and exciting application of GTNs: accelerating the evaluation of candidate architectures for neural architecture search (NAS), which is rate-limited by such evaluations, enabling massive speed-ups in NAS. GTN-NAS improves the NAS state of the art, finding higher performing architectures when controlling for the search proposal mechanism. GTN-NAS also is competitive with the overall state of the art approaches, which achieve top performance while using orders of magnitude less computation than typical NAS methods. Speculating forward, GTNs may represent a first step toward the ambitious goal of algorithms that generate their own training data and, in doing so, open a variety of interesting new research questions and directions.

References in corpus (16)

• Adam: A Method for Stochastic Optimization
• Distilling the Knowledge in a Neural Network
• Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks
• Neural Architecture Search with Reinforcement Learning
• Large Scale GAN Training for High Fidelity Natural Image Synthesis
• Improved Regularization of Convolutional Neural Networks with Cutout
• Improved Techniques for Training GANs
• Evolution Strategies as a Scalable Alternative to Reinforcement Learning
• Neural Architecture Optimization
• Learning Dexterous In-Hand Manipulation
• Learning Active Learning from Data
• Paired Open-Ended Trailblazer (POET): Endlessly Generating Increasingly Complex and Diverse Learning Environments and Their Solutions
• Graph HyperNetworks for Neural Architecture Search
• Meta-Learning and Universality: Deep Representations and Gradient Descent can Approximate any Learning Algorithm
• Meta-Learning Update Rules for Unsupervised Representation Learning
• AI-GAs: AI-generating algorithms, an alternate paradigm for producing general artificial intelligence

Cited by in corpus (3)

• Weight-Sharing Neural Architecture Search: A Battle to Shrink the Optimization Gap
• Meta Back-translation
• A study of traits that affect learnability in GANs