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
Proceedings ArticleDOI
09 Jul 2020
TL;DR: This paper proposes a novel grouped operation dropout algorithm named DropNAS to fix the problems with DARTS and demonstrates that DropNAS solves the above issues and achieves promising performance.
Abstract: Neural architecture search (NAS) has shown encouraging results in automating the architecture design. Recently, DARTS relaxes the search process with a differentiable formulation that leverages weight-sharing and SGD for cost reduction of NAS. In DARTS, all candidate operations are trained simultaneously during the network weight training step. Our empirical results show that this training procedure leads to the co-adaption problem and Matthew Effect: operations with fewer parameters would be trained maturely earlier. This causes two problems: firstly, the operations with more parameters may never have the chance to express the desired function since those with less have already done the job; secondly, the system will punish those underperforming operations by lowering their architecture parameter and backward smaller loss gradients, this causes the Matthew Effect. In this paper, we systematically study these problems and propose a novel grouped operation dropout algorithm named DropNAS to fix the problems with DARTS. Extensive experiments demonstrate that DropNAS solves the above issues and achieves promising performance. Specifically, DropNAS achieves 2.26% test error on CIFAR-10, 16.39% on CIFAR-100 and 23.4% on ImageNet (with the same training hyperparameters as DARTS for a fair comparison). It is also observed that DropNAS is robust across variants of the DARTS search space. Code is available at https://github.com/wiljohnhong/DropNAS.

38 citations

Posted Content
TL;DR: In this article, the authors study two procedures (reverse-mode and forward-mode) for computing the gradient of the validation error with respect to the hyperparameters of any iterative learning algorithm such as stochastic gradient descent.
Abstract: We study two procedures (reverse-mode and forward-mode) for computing the gradient of the validation error with respect to the hyperparameters of any iterative learning algorithm such as stochastic gradient descent. These procedures mirror two methods of computing gradients for recurrent neural networks and have different trade-offs in terms of running time and space requirements. Our formulation of the reverse-mode procedure is linked to previous work by Maclaurin et al. [2015] but does not require reversible dynamics. The forward-mode procedure is suitable for real-time hyperparameter updates, which may significantly speed up hyperparameter optimization on large datasets. We present experiments on data cleaning and on learning task interactions. We also present one large-scale experiment where the use of previous gradient-based methods would be prohibitive.

38 citations

Proceedings ArticleDOI
15 Oct 2018
TL;DR: This work proposes an efficient greedy neural architecture search approach (GNAS) to automatically discover the optimal tree-like deep architecture for multi-attribute learning and demonstrates the effectiveness and compactness of neural architectures derived by GNAS.
Abstract: A key problem in deep multi-attribute learning is to effectively discover the inter-attribute correlation structures. Typically, the conventional deep multi-attribute learning approaches follow the pipeline of manually designing the network architectures based on task-specific expertise prior knowledge and careful network tunings, leading to the inflexibility for various complicated scenarios in practice. Motivated by addressing this problem, we propose an efficient greedy neural architecture search approach (GNAS) to automatically discover the optimal tree-like deep architecture for multi-attribute learning. In a greedy manner, GNAS divides the optimization of global architecture into the optimizations of individual connections step by step. By iteratively updating the local architectures, the global tree-like architecture gets converged where the bottom layers are shared across relevant attributes and the branches in top layers more encode attribute-specific features. Experiments on three benchmark multi-attribute datasets show the effectiveness and compactness of neural architectures derived by GNAS, and also demonstrate the efficiency of GNAS in searching neural architectures.

38 citations


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

  • ...Typically, Zoph and Le [29] use 800 GPUs and 28 days to discover the convolutional architecture on Cifar-10 dataset by exploring 12,800 individual architectures....

    [...]

  • ...Œe recently proposed neural architecture search (NAS) [29, 30] employs an RNN controller to sample candidate architectures and updating the controller under the guidance of performances of sampled architectures....

    [...]

  • ...• GNAS is a non-parametric approach that it refrains from the loop of adopting extra parameters for meta-learning (such as Bayesian optimization (BO) [23] and reinforcement learning (RL) [29, 30])....

    [...]

  • ...A variety of approaches including random search [3], Bayesian optimization [11, 16, 23], evolutionary algorithm [21], and reinforcement learning [20, 29] are proposed for neural architecture optimization....

    [...]

Proceedings ArticleDOI
01 Jan 2020
TL;DR: A novel framework to co-explore NAS space and NoC Design Search (NDS) space with the objective to maximize accuracy and throughput is proposed and a multi-phase manager is developed to guide NANDS to gradually converge to solutions with the best accuracy-throughput tradeoff.
Abstract: Hardware-aware Neural Architecture Search (NAS), which automatically finds an architecture that works best on a given hardware design, has prevailed in response to the ever-growing demand for real-time Artificial Intelligence (AI). However, in many situations, the underlying hardware is not pre-determined. We argue that simply assuming an arbitrary yet fixed hardware design will lead to inferior solutions, and it is best to co-explore neural architecture space and hardware design space for the best pair of neural architecture and hardware design. To demonstrate this, we employ Network-on-Chip (NoC) as the infrastructure and propose a novel framework, namely NANDS, to co-explore NAS space and NoC Design Search (NDS) space with the objective to maximize accuracy and throughput. Since two metrics are tightly coupled, we develop a multi-phase manager to guide NANDS to gradually converge to solutions with the best accuracy-throughput tradeoff. On top of it, we propose techniques to detect and alleviate timing performance bottleneck, which allows better and more efficient exploration of NDS space. Experimental results on common datasets, CIFAR10, CIFAR-100 and STL-10, show that compared with state-of-the-art hardware-aware NAS, NANDS can achieve 42.99% higher throughput along with 1.58% accuracy improvement. There are cases where hardware-aware NAS cannot find any feasible solutions while NANDS can.

38 citations


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

  • ...Design Space Exploration: In this set of experiment, we compare NANDS framework on CIFAR-10 with the original NAS framework [1], and the state-of-the-art HW-aware NAS [20]....

    [...]

  • ...Unlike the implementation in [1] with monocriteria (accuracy), NAS controller in NANDS will take both accuracy and throughput to update RNN....

    [...]

  • ...For NAS [1] and HW-aware NAS [20], we report the finally identified architectures....

    [...]

  • ...For instance, as shown in Figure 2(a), for the neural network with 15 layers obtained by NAS in [1], we observe that the generated architecture (1) contains up to 8 different types of kernels leading the use of a uniform design to be inefficient; (2) involves a lot of skip connections between layers resulting in a large amount of data movement....

    [...]

  • ...Specifically, we use a reinforcement learningbased NAS controller [1, 6] as the backbone to explore NAS space....

    [...]

Posted Content
TL;DR: It is shown that the expressiveness of Softmax-based models (including the majority of neural language models) is limited by a Softmax bottleneck, and a simple and effective method is proposed to address this issue.
Abstract: We formulate language modeling as a matrix factorization problem, and show that the expressiveness of Softmax-based models (including the majority of neural language models) is limited by a Softmax bottleneck. Given that natural language is highly context-dependent, this further implies that in practice Softmax with distributed word embeddings does not have enough capacity to model natural language. We propose a simple and effective method to address this issue, and improve the state-of-the-art perplexities on Penn Treebank and WikiText-2 to 47.69 and 40.68 respectively. The proposed method also excels on the large-scale 1B Word dataset, outperforming the baseline by over 5.6 points in perplexity.

38 citations

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)....

    [...]