Progressive Neural Architecture Search
Explore this paper's citation graph
Summary
This work proposes a new method for learning the structure of convolutional neural networks (CNNs) that is more efficient than recent state-of-the-art methods based on reinforcement learning and evolutionary algorithms using a sequential model-based optimization (SMBO) strategy.
- Type
- preprint
- Published
- 2017-12-02
- Cited by
- 2,167
- References
- 53
- Access
- Open access
- OpenAlex
- https://openalex.org/W2771727678
- Semantic Scholar
- https://api.semanticscholar.org/CorpusID:40430109
Keywords
Reinforcement learning, Computer science, Artificial intelligence, Convolutional neural network, State (computer science)
References
- Very Deep Convolutional Networks for Large-Scale Image Recognition
- Going deeper with convolutions
- ImageNet: A large-scale hierarchical image database
- Evolving Neural Networks through Augmenting Topologies
- Practical Bayesian Optimization of Machine Learning Algorithms
- ImageNet classification with deep convolutional neural networks
- Rethinking the Inception Architecture for Computer Vision
- Taking the Human Out of the Loop: A Review of Bayesian Optimization
- Deep Residual Learning for Image Recognition
- On a Family of Decomposable Kernels on Sequences
- Speeding Up Automatic Hyperparameter Optimization of Deep Neural Networks by Extrapolation of Learning Curves
- SGDR: Stochastic Gradient Descent with Warm Restarts
- Aggregated Residual Transformations for Deep Neural Networks
- Designing Neural Network Architectures using Reinforcement Learning
- Towards Automatically-Tuned Neural Networks
- Large-Scale Evolution of Image Classifiers
- DeepArchitect: Automatically Designing and Training Deep Architectures
- MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications
- Dual Path Networks
- Reinforcement Learning for Architecture Search by Network Transformation
Cited by
- A genetic programming approach to designing convolutional neural network architectures
- Lifelong Generative Modeling
- Squeeze-and-Excitation Networks
- Deep Expander Networks: Efficient Deep Networks from Graph Theory
- Regularized Evolution for Image Classifier Architecture Search
- Faster Discovery of Neural Architectures by Searching for Paths in a Large Model
- Fast Neural Architecture Construction using EnvelopeNets
- Efficient image dataset classification difficulty estimation for predicting deep-learning accuracy
- Evolving Hierarchical Structures for Convolutional Neural Networks using Jagged Arrays
- Network Transplanting
- Multi-objective Architecture Search for CNNs
- PANDA: Facilitating Usable AI Development
- Path-Level Network Transformation for Efficient Architecture Search
- Parallel Architecture and Hyperparameter Search via Successive Halving and Classification
- AlphaX: eXploring Neural Architectures with Deep Neural Networks and Monte Carlo Tree Search
- AutoAugment: Learning Augmentation Policies from Data
- Deep Convolutional Neural Networks for Map-Type Classification
- Heterogeneous transfer learning
- Auto-Meta: Automated Gradient Based Meta Learner Search
- Towards Automated Deep Learning: Efficient Joint Neural Architecture and Hyperparameter Search
Related papers
- Cost-Efficient Reinforcement Learning for Optimal Trade Execution on Dynamic Market Environment
- Performance of distributed multi-agent multi-state reinforcement spectrum management using different exploration schemes
- Finding shortcuts from episode in multi-agent reinforcement learning
- Spatially and Seamlessly Hierarchical Reinforcement Learning for State Space and Policy space in Autonomous Driving
- Hierarchical Reinforcement Learning Based on System Model
- A reinforcement learning with adaptive state space recruitment strategy for real autonomous mobile robots