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 2022
TL;DR: A novel multi-objective evolutionary NAS algorithm is proposed to optimally design multi-layered feed-forward ANNs by balancing the aspects of parsimony and accuracy and provides a generic method applicable to any kind of data/model from process industries.
Abstract: State-of-the-art infrastructure, excellent computational facilities and ubiquitous connectivity across the industries have led to the generation of large amounts of heterogeneous process data. At the same time, the applicability of machine learning and artificial intelligence is witnessing a significant rise in academics and engineering, leading to the development of a large number of resources and tools. However, the number of research works and applications aimed at implementing data sciences to problems in process industries is far less. The proposed work aims to fill the niche by proposing Artificial Neural Network (ANN)-based surrogate construction using extremely nonlinear, static, high dimensional (32 features) noisy data sampled irregularly from inlet and outlet streams of hot rolling process in iron and steel making industry. Though ANNs are used extensively for modelling nonlinear data, literature survey has shown that their modelling is governed by heuristics thus making them inefficient for use in process industries. This aspect is of high relevance in contemporary times as hyper-parameter optimization, automated machine learning and neural architecture search (NAS) constitute a major share of current research in data sciences. We propose a novel multi-objective evolutionary NAS algorithm to optimally design multi-layered feed-forward ANNs by balancing the aspects of parsimony and accuracy. The integer nonlinear programming problem of ANN design is solved using binary coded Non-Dominated Sorting Genetic Algorithm (NSGA-II). ANNs designed for the hot rolling process are found to demonstrate an accuracy of 0.98 (averaged on three outputs) measured in terms of correlation coefficient R2 on the test set. The successful construction of accurate and optimal ANNs provides a first-of-its-kind model for the hot rolling process in the iron and steel making industry. The proposed method can minimize the chances of over-fitting in ANNs and provides a generic method applicable to any kind of data/model from process industries.

1 citations

Book ChapterDOI
Martin Wistuba1
14 Sep 2020
TL;DR: In this paper, the authors propose a general applicable framework that introduces only minor changes to existing optimizers to leverage knowledge from previous searches on different tasks and demonstrate the complexity of the integration of the framework as well as its impact.
Abstract: The term Neural Architecture Search (NAS) refers to the automatic optimization of network architectures for a new, previously unknown task. Since testing an architecture is computationally very expensive, many optimizers need days or even weeks to find suitable architectures. However, this search time can be significantly reduced if knowledge from previous searches on different tasks is reused. In this work, we propose a generally applicable framework that introduces only minor changes to existing optimizers to leverage this feature. As an example, we select an existing optimizer and demonstrate the complexity of the integration of the framework as well as its impact. In experiments on CIFAR-10 and CIFAR-100, we observe a reduction in the search time from 200 to only 6 GPU days, a speed up by a factor of 33. In addition, we observe new records of 1.99 and 14.06 for NAS optimizers on the CIFAR benchmarks, respectively. In a separate study, we analyze the impact of the amount of source and target data. Empirically, we demonstrate that the proposed framework generally gives better results and, in the worst case, is just as good as the unmodified optimizer.

1 citations

Posted Content
TL;DR: A Reinforcement Learning (RL) policy-based search algorithm with a direct objective of finding transferable layer-wise pruning profiles using many models for the same architecture is developed and results show that the transferred RL-based profiles are as good or better than best profiles found on the original dataset via exhaustive search.
Abstract: In the last decade convolutional neural networks have become gargantuan. Pre-trained models, when used as initializers are able to fine-tune ever larger networks on small datasets. Consequently, not all the convolutional features that these fine-tuned models detect are requisite for the end-task. Several works of channel pruning have been proposed to prune away compute and memory from models that were trained already. Typically, these involve policies that decide which and how many channels to remove from each layer leading to channel-wise and/or layer-wise pruning profiles, respectively. In this paper, we conduct several baseline experiments and establish that profiles from random channel-wise pruning policies are as good as metric-based ones. We also establish that there may exist profiles from some layer-wise pruning policies that are measurably better than common baselines. We then demonstrate that the top layer-wise pruning profiles found using an exhaustive random search from one datatset are also among the top profiles for other datasets. This implies that we could identify out-of-the-box layer-wise pruning profiles using benchmark datasets and use these directly for new datasets. Furthermore, we develop a Reinforcement Learning (RL) policy-based search algorithm with a direct objective of finding transferable layer-wise pruning profiles using many models for the same architecture. We use a novel reward formulation that drives this RL search towards an expected compression while maximizing accuracy. Our results show that our transferred RL-based profiles are as good or better than best profiles found on the original dataset via exhaustive search. We then demonstrate that if we found the profiles using a mid-sized dataset such as Cifar10/100, we are able to transfer them to even a large dataset such as Imagenet.

1 citations


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

  • ...Recently, RL has emerged as a powerful and general approach in various domains from complex games [5, 41, 45, 47] to simulated robotics tasks [2, 18, 37], all the way to neural architecture search [7, 35, 58]....

    [...]

  • ...To the best of our knowledge, this is the first work that performs generalizable and transferable channel pruning with (either a random search or) a RL-based approach, even though there are some prior works that train RL-based policies for pruning a particular network [3, 6, 14,58]....

    [...]

Proceedings ArticleDOI
18 Jul 2022
TL;DR: The experimental results demonstrate that ERNAS can be trained effectively enough with extremely limited training data and the accuracy of the neural architecture search result produced by ERNAS is greater than that of the SOTA methods.
Abstract: Search space, searching method, and candidate evaluation scheme are critical to the success of Neural Architecture Search (NAS), especially the evaluation strategy. An effective and efficient neural architecture performance evaluator could successfully save computing costs and search time while guiding the NAS process to the optimal solution as fast as possible. Most existing NAS algorithms attempt to compute the absolute accuracy of the candidate architecture, which is almost impossible to achieve and meaningless to the final performance improvement. In this paper, we propose ERNAS, a novel neural architecture performance evaluation approach that optimizes the ranking of the candidate architecture performance, rather than the absolute accuracy itself. With the help of ERNAS, many existing NAS methods could achieve better performance without further evaluation. The experimental results demonstrate that ERNAS can be trained effectively enough with extremely limited training data (423 neural architectures randomly sampled form NAS-Bench-101, which is only 0.1% of the entire search space). The accuracy of the neural architecture search result produced by ERNAS is greater than that of the SOTA methods.

1 citations

Journal ArticleDOI
TL;DR: This work proposes a architecture that leverages advances in probabilistic neural architecture search and approximate Bayesian inference to generate ensembles form the joint distribution of neural network architectures and weights that showed improvement both with in-dist distribution and out-of-distribution.
Abstract: Robust machine learning models with accurately calibrated uncertainties are crucial for safety-critical applications. Probabilistic machine learning and especially the Bayesian formalism provide a systematic framework to incorporate robustness through the distributional estimates and reason about uncertainty. Recent works have shown that approximate inference approaches that take the weight space uncertainty of neural networks to generate ensemble prediction are the state-of-the-art. However, architecture choices have mostly been ad hoc, which essentially ignores the epistemic uncertainty from the architecture space. To this end, we propose a Unified probabilistic architecture and weight ensembling Neural Architecture Search (UraeNAS) that leverages advances in probabilistic neural architecture search and approximate Bayesian inference to generate ensembles form the joint distribution of neural network architectures and weights. The proposed approach showed a significant improvement both with in-distribution (0.86% in accuracy, 42% in ECE) CIFAR-10 and out-of-distribution (2.43% in accuracy, 30% in ECE) CIFAR-10-C compared to the baseline deterministic approach.

1 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)....

    [...]