scispace - formally typeset
Search or ask a question
Posted Content

Neural Architecture Search with Reinforcement Learning

Barret Zoph1, Quoc V. Le1
05 Nov 2016-arXiv: Learning-
TL;DR: This paper uses a recurrent network to generate the model descriptions of neural networks and trains this RNN with reinforcement learning to maximize the expected accuracy of the generated architectures on a validation set.
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.
Citations
More filters
Book ChapterDOI
01 Jan 2019
TL;DR: This chapter proposes a large-scale systematic experimental setup in order to design and evaluate neural architectures for HAR applications and uses a Bayesian optimization procedure based on a Gaussian process surrogate model to tune architectures’ hyper-parameters.
Abstract: Design of neural architectures is a critical aspect in deep-learning based methods. In this chapter, we explore the suitability of different neural architectures for the recognition of mobility-related human activities. Neural architecture search (NAS) is getting a lot of attention in the machine learning community and improves deep learning models’ performances in many tasks like language modeling and image recognition. Deep learning techniques were successfully applied to human activity recognition (HAR). However, the design of competitive architectures remains cumbersome, time-consuming, and rely strongly on domain expertise. To address this, we propose a large-scale systematic experimental setup in order to design and evaluate neural architectures for HAR applications. Specifically, we use a Bayesian optimization (BO) procedure based on a Gaussian process surrogate model in order to tune architectures’ hyper-parameters. We train and evaluate more than 600 different architectures which are then analyzed via the functional ANalysis Of VAriance (fANOVA) framework to assess hyper-parameters relevance. We experiment our approach on the Sussex-Huawei Locomotion and Transportation (SHL) dataset, a highly versatile, sensor-rich and precisely annotated dataset of human locomotion modes.

3 citations

Proceedings Article
TL;DR: In this article , task model parameters are structured into multiple subspaces, and each subspace represents one type of meta-knowledge, and an algorithm is proposed to learn the meta-parameters (i.e., subspace bases).
Abstract: Meta-learning aims to extract meta-knowledge from historical tasks to accelerate learning on new tasks. Typical meta-learning algorithms like MAML learn a globally-shared meta-model for all tasks. However, when the task environments are complex, task model parameters are diverse and a common meta-model is insufficient to capture all the meta-knowledge. To address this challenge, in this paper, task model parameters are structured into multiple subspaces, and each subspace represents one type of meta-knowledge. We propose an algorithm to learn the meta-parameters (i.e., subspace bases). We theoretically study the generalization properties of the learned subspaces. Experiments on regression and classification meta-learning datasets verify the effectiveness of the proposed algorithm.

3 citations

Book ChapterDOI
17 Sep 2019
TL;DR: This paper proposes the introduction of a penalty term to control the model complexity of obtained structures using the number of weights or units and derive its analytical natural gradient based on the stochastic gradient descent.
Abstract: A method of simultaneously optimizing both the structure of neural networks and the connection weights in a single training loop can reduce the enormous computational cost of neural architecture search. We focus on the probabilistic model-based dynamic neural network structure optimization that considers the probability distribution of structure parameters and simultaneously optimizes both the distribution parameters and connection weights based on gradient methods. Since the existing algorithm searches for the structures that only minimize the training loss, this method might find overly complicated structures. In this paper, we propose the introduction of a penalty term to control the model complexity of obtained structures. We formulate a penalty term using the number of weights or units and derive its analytical natural gradient. The proposed method minimizes the objective function injected the penalty term based on the stochastic gradient descent. We apply the proposed method in the unit selection of a fully-connected neural network and the connection selection of a convolutional neural network. The experimental results show that the proposed method can control model complexity while maintaining performance.

3 citations


Cites background from "Neural Architecture Search with Rei..."

  • ...Another approach trains the neural network that generates the network architecture using policy gradient-based reinforcement learning methods [23]....

    [...]

  • ...However, these approaches require huge computational resources; several works conducted the experiments using more than 100 GPUs [15,23], as the evaluation of a candidate structure requires model training and takes several hours in the case of DNNs....

    [...]

Posted Content
TL;DR: This work proposes Intra-Ensemble, an end-to-end ensemble strategy with stochastic channel recombination operations to train several sub-networks simultaneously within one neural network, which enhances ensemble performance.
Abstract: Improving model performance is always the key problem in machine learning including deep learning. However, stand-alone neural networks always suffer from marginal effect when stacking more layers. At the same time, ensemble is an useful technique to further enhance model performance. Nevertheless, training several independent deep neural networks for ensemble costs multiple resources. If so, is it possible to utilize ensemble in only one neural network? In this work, we propose Intra-Ensemble, an end-to-end ensemble strategy with stochastic channel recombination operations to train several sub-networks simultaneously within one neural network. Additional parameter size is marginal since the majority of parameters are mutually shared. Meanwhile, stochastic channel recombination significantly increases the diversity of sub-networks, which finally enhances ensemble performance. Extensive experiments and ablation studies prove the applicability of intra-ensemble on various kinds of datasets and network architectures.

3 citations


Cites background from "Neural Architecture Search with Rei..."

  • ...Neural architecture search Recently, neural architecture search(NAS) [24, 25, 26] has shown great improvement in designing smaller and more accurate network architectures....

    [...]

Proceedings ArticleDOI
01 Nov 2019
TL;DR: A new deep reinforcement learning architecture with model-based acceleration to optimize hyperparameters for any machine learning model, constructed by a Long Short-Term Memory Network, which outperforms random search, Bayesian optimization, and Tree-structured Parzen Estimator in terms of accuracy, time efficiency and stability.
Abstract: Hyperparameter optimization is a key part of AutoML. In recent years, there have been successful hyperparameter optimization algorithms. However, these methods still face several challenges, such as high cost of evaluating large models or large datasets. In this paper, we introduce a new deep reinforcement learning architecture with model-based acceleration to optimize hyperparameters for any machine learning model. In this method, an agent constructed by a Long Short-Term Memory Network aims at maximizing the expected accuracy of a machine learning model on a validation set. To speed up training, we employ a model to predict the accuracy on a validation set instead of evaluating a machine learning model. To effectively train the agent and the predictive model, Real-Predictive-Real training process is proposed. Besides, to reduce the variance, we propose a bootstrap pool to guide the exploration in the search space. The experiment was carried out by optimizing hyperparameters of two widely used machine learning models: Random Forests and XGBoost. Experimental results show that the proposed method outperforms random search, Bayesian optimization, and Tree-structured Parzen Estimator in terms of accuracy, time efficiency and stability.

3 citations


Cites background from "Neural Architecture Search with Rei..."

  • ...A work related to ours is [22] proposed by Google Brain team in early 2017....

    [...]

  • ...It is noted that the proposed model is different from the one in [22] since:...

    [...]

References
More filters
Proceedings ArticleDOI
27 Jun 2016
TL;DR: In this article, the authors proposed a residual learning framework to ease the training of networks that are substantially deeper than those used previously, which won the 1st place on the ILSVRC 2015 classification task.
Abstract: Deeper neural networks are more difficult to train. We present a residual learning framework to ease the training of networks that are substantially deeper than those used previously. We explicitly reformulate the layers as learning residual functions with reference to the layer inputs, instead of learning unreferenced functions. We provide comprehensive empirical evidence showing that these residual networks are easier to optimize, and can gain accuracy from considerably increased depth. On the ImageNet dataset we evaluate residual nets with a depth of up to 152 layers—8× deeper than VGG nets [40] but still having lower complexity. An ensemble of these residual nets achieves 3.57% error on the ImageNet test set. This result won the 1st place on the ILSVRC 2015 classification task. We also present analysis on CIFAR-10 with 100 and 1000 layers. The depth of representations is of central importance for many visual recognition tasks. Solely due to our extremely deep representations, we obtain a 28% relative improvement on the COCO object detection dataset. Deep residual nets are foundations of our submissions to ILSVRC & COCO 2015 competitions1, where we also won the 1st places on the tasks of ImageNet detection, ImageNet localization, COCO detection, and COCO segmentation.

123,388 citations

Proceedings Article
01 Jan 2015
TL;DR: This work introduces Adam, an algorithm for first-order gradient-based optimization of stochastic objective functions, based on adaptive estimates of lower-order moments, and provides a regret bound on the convergence rate that is comparable to the best known results under the online convex optimization framework.
Abstract: We introduce Adam, an algorithm for first-order gradient-based optimization of stochastic objective functions, based on adaptive estimates of lower-order moments. The method is straightforward to implement, is computationally efficient, has little memory requirements, is invariant to diagonal rescaling of the gradients, and is well suited for problems that are large in terms of data and/or parameters. The method is also appropriate for non-stationary objectives and problems with very noisy and/or sparse gradients. The hyper-parameters have intuitive interpretations and typically require little tuning. Some connections to related algorithms, on which Adam was inspired, are discussed. We also analyze the theoretical convergence properties of the algorithm and provide a regret bound on the convergence rate that is comparable to the best known results under the online convex optimization framework. Empirical results demonstrate that Adam works well in practice and compares favorably to other stochastic optimization methods. Finally, we discuss AdaMax, a variant of Adam based on the infinity norm.

111,197 citations

Proceedings Article
04 Sep 2014
TL;DR: This work investigates the effect of the convolutional network depth on its accuracy in the large-scale image recognition setting using an architecture with very small convolution filters, which shows that a significant improvement on the prior-art configurations can be achieved by pushing the depth to 16-19 weight layers.
Abstract: In this work we investigate the effect of the convolutional network depth on its accuracy in the large-scale image recognition setting. Our main contribution is a thorough evaluation of networks of increasing depth using an architecture with very small (3x3) convolution filters, which shows that a significant improvement on the prior-art configurations can be achieved by pushing the depth to 16-19 weight layers. These findings were the basis of our ImageNet Challenge 2014 submission, where our team secured the first and the second places in the localisation and classification tracks respectively. We also show that our representations generalise well to other datasets, where they achieve state-of-the-art results. We have made our two best-performing ConvNet models publicly available to facilitate further research on the use of deep visual representations in computer vision.

55,235 citations


"Neural Architecture Search with Rei..." refers methods in this paper

  • ...Along with this success is a paradigm shift from feature designing to architecture designing, i.e., from SIFT (Lowe, 1999), and HOG (Dalal & Triggs, 2005), to AlexNet (Krizhevsky et al., 2012), VGGNet (Simonyan & Zisserman, 2014), GoogleNet (Szegedy et al., 2015), and ResNet (He et al., 2016a)....

    [...]

Journal ArticleDOI
01 Jan 1998
TL;DR: In this article, a graph transformer network (GTN) is proposed for handwritten character recognition, which can be used to synthesize a complex decision surface that can classify high-dimensional patterns, such as handwritten characters.
Abstract: Multilayer neural networks trained with the back-propagation algorithm constitute the best example of a successful gradient based learning technique. Given an appropriate network architecture, gradient-based learning algorithms can be used to synthesize a complex decision surface that can classify high-dimensional patterns, such as handwritten characters, with minimal preprocessing. This paper reviews various methods applied to handwritten character recognition and compares them on a standard handwritten digit recognition task. Convolutional neural networks, which are specifically designed to deal with the variability of 2D shapes, are shown to outperform all other techniques. Real-life document recognition systems are composed of multiple modules including field extraction, segmentation recognition, and language modeling. A new learning paradigm, called graph transformer networks (GTN), allows such multimodule systems to be trained globally using gradient-based methods so as to minimize an overall performance measure. Two systems for online handwriting recognition are described. Experiments demonstrate the advantage of global training, and the flexibility of graph transformer networks. A graph transformer network for reading a bank cheque is also described. It uses convolutional neural network character recognizers combined with global training techniques to provide record accuracy on business and personal cheques. It is deployed commercially and reads several million cheques per day.

42,067 citations

Proceedings ArticleDOI
20 Jun 2005
TL;DR: It is shown experimentally that grids of histograms of oriented gradient (HOG) descriptors significantly outperform existing feature sets for human detection, and the influence of each stage of the computation on performance is studied.
Abstract: We study the question of feature sets for robust visual object recognition; adopting linear SVM based human detection as a test case. After reviewing existing edge and gradient based descriptors, we show experimentally that grids of histograms of oriented gradient (HOG) descriptors significantly outperform existing feature sets for human detection. We study the influence of each stage of the computation on performance, concluding that fine-scale gradients, fine orientation binning, relatively coarse spatial binning, and high-quality local contrast normalization in overlapping descriptor blocks are all important for good results. The new approach gives near-perfect separation on the original MIT pedestrian database, so we introduce a more challenging dataset containing over 1800 annotated human images with a large range of pose variations and backgrounds.

31,952 citations


"Neural Architecture Search with Rei..." refers methods in this paper

  • ...Along with this success is a paradigm shift from feature designing to architecture designing, i.e., from SIFT (Lowe, 1999), and HOG (Dalal & Triggs, 2005), to AlexNet (Krizhevsky et al., 2012), VGGNet (Simonyan & Zisserman, 2014), GoogleNet (Szegedy et al., 2015), and ResNet (He et al., 2016a)....

    [...]