Posted Content•
Neural Architecture Search with Reinforcement 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
Posted Content•
TL;DR: DNW provides an effective mechanism for discovering sparse subnetworks of predefined architectures in a single training run and is regarded as unifying core aspects of the neural architecture search problem with sparse neural network learning.
Abstract: The success of neural networks has driven a shift in focus from feature engineering to architecture engineering. However, successful networks today are constructed using a small and manually defined set of building blocks. Even in methods of neural architecture search (NAS) the network connectivity patterns are largely constrained. In this work we propose a method for discovering neural wirings. We relax the typical notion of layers and instead enable channels to form connections independent of each other. This allows for a much larger space of possible networks. The wiring of our network is not fixed during training -- as we learn the network parameters we also learn the structure itself. Our experiments demonstrate that our learned connectivity outperforms hand engineered and randomly wired networks. By learning the connectivity of MobileNetV1we boost the ImageNet accuracy by 10% at ~41M FLOPs. Moreover, we show that our method generalizes to recurrent and continuous time networks. Our work may also be regarded as unifying core aspects of the neural architecture search problem with sparse neural network learning. As NAS becomes more fine grained, finding a good architecture is akin to finding a sparse subnetwork of the complete graph. Accordingly, DNW provides an effective mechanism for discovering sparse subnetworks of predefined architectures in a single training run. Though we only ever use a small percentage of the weights during the forward pass, we still play the so-called initialization lottery with a combinatorial number of subnetworks. Code and pretrained models are available at this https URL while additional visualizations may be found at this https URL.
75 citations
Cites background from "Neural Architecture Search with Rei..."
...A myriad of recent efforts attempt to automate the process of the architecture design by searching among a set of smaller well-known building blocks [30, 34, 37, 19, 2, 20]....
[...]
Posted Content•
TL;DR: This paper incorporates a variant of multi-objective genetic algorithm NSGA-II, in which the search space is composed of various cells so that crossovers and mutations can be performed at the cell level, and prevents the searched models from degrading during the evolution process.
Abstract: Fabricating neural models for a wide range of mobile devices demands for a specific design of networks due to highly constrained resources. Both evolution algorithms (EA) and reinforced learning methods (RL) have been dedicated to solve neural architecture search problems. However, these combinations usually concentrate on a single objective such as the error rate of image classification. They also fail to harness the very benefits from both sides. In this paper, we present a new multi-objective oriented algorithm called MoreMNAS (Multi-Objective Reinforced Evolution in Mobile Neural Architecture Search) by leveraging good virtues from both EA and RL. In particular, we incorporate a variant of multi-objective genetic algorithm NSGA-II, in which the search space is composed of various cells so that crossovers and mutations can be performed at the cell level. Moreover, reinforced control is mixed with a natural mutating process to regulate arbitrary mutation, maintaining a delicate balance between exploration and exploitation. Therefore, not only does our method prevent the searched models from degrading during the evolution process, but it also makes better use of learned knowledge. Our experiments conducted in Super-resolution domain (SR) deliver rivalling models compared to some state-of-the-art methods with fewer FLOPS.
75 citations
Cites background or methods from "Neural Architecture Search with Rei..."
...It is represented as a sequence of parameters to describe raw layers in NAS (Zoph & Le, 2016), followed by MetaQNN (Baker et al., 2017), ENAS (Pham et al., 2018), in which the selection of parameters is essentially finding subgraphs in a single directed acyclic graph (DAG)....
[...]
...Pure reinforced methods was initiated by NAS (Zoph & Le, 2016), later echoed by NASNet (Zoph et al., 2017), ENAS (Pham et al., 2018), MetaQNN (Baker et al., 2017), MnasNet (Tan et al., 2018), MONAS (Hsu et al., 2018) etc....
[...]
...In fact, macro-level search (Zoph & Le, 2016) is harmful to mobile devices regarding some underlying hardware designs (Ma et al., 2018)....
[...]
...…methods feature a recurrent neural controller (RNN or LSTM) to generate models, with its parameters updated by a family of Policy Gradient algorithms: REINFORCEMENT (Zoph & Le, 2016; Pham et al., 2018; Hsu et al., 2018), Proximal Policy Optimization (Zoph et al., 2017; Tan et al., 2018)....
[...]
Posted Content•
TL;DR: In this article, the authors proposed AdaNet, which adaptively learns both the structure of the network and its weights based on a solid theoretical analysis, including data-dependent generalization guarantees that they prove and discuss in detail.
Abstract: We present new algorithms for adaptively learning artificial neural networks. Our algorithms (AdaNet) adaptively learn both the structure of the network and its weights. They are based on a solid theoretical analysis, including data-dependent generalization guarantees that we prove and discuss in detail. We report the results of large-scale experiments with one of our algorithms on several binary classification tasks extracted from the CIFAR-10 dataset. The results demonstrate that our algorithm can automatically learn network structures with very competitive performance accuracies when compared with those achieved for neural networks found by standard approaches.
74 citations
01 Jul 2018
TL;DR: This work takes a fine-grained look at the different architectures for NMT and introduces an Architecture Definition Language (ADL) allowing for a flexible combination of common building blocks and shows that self-attention is much more important on the encoder side than on the decoder side.
Abstract: With recent advances in network architectures for Neural Machine Translation (NMT) recurrent models have effectively been replaced by either convolutional or self-attentional approaches, such as in the Transformer. While the main innovation of the Transformer architecture is its use of self-attentional layers, there are several other aspects, such as attention with multiple heads and the use of many attention layers, that distinguish the model from previous baselines. In this work we take a fine-grained look at the different architectures for NMT. We introduce an Architecture Definition Language (ADL) allowing for a flexible combination of common building blocks. Making use of this language we show in experiments that one can bring recurrent and convolutional models very close to the Transformer performance by borrowing concepts from the Transformer architecture, but not using self-attention. Additionally, we find that self-attention is much more important on the encoder side than on the decoder side, where it can be replaced by a RNN or CNN without a loss in performance in most settings. Surprisingly, even a model without any target side self-attention performs well.
74 citations
Cites background from "Neural Architecture Search with Rei..."
...For the application of image classification there have been several recent successful efforts of automatically searching for successful architectures (Zoph and Le, 2016; Negrinho and Gordon, 2017; Liu et al., 2017)....
[...]
TL;DR: This work suggests using a lower-dimensional representation of the original data to quickly identify promising areas in the hyperparameter space, which can be used to initialize the optimization algorithm for the original, higher-dimensional data.
Abstract: Most learning algorithms require the practitioner to manually set the values of many hyperparameters before the learning process can begin. However, with modern algorithms, the evaluation of a given hyperparameter setting can take a considerable amount of time and the search space is often very high-dimensional. We suggest using a lower-dimensional representation of the original data to quickly identify promising areas in the hyperparameter space. This information can then be used to initialize the optimization algorithm for the original, higher-dimensional data. We compare this approach with the standard procedure of optimizing the hyperparameters only on the original input.
We perform experiments with various state-of-the-art hyperparameter optimization algorithms such as random search, the tree of parzen estimators (TPEs), sequential model-based algorithm configuration (SMAC), and a genetic algorithm (GA). Our experiments indicate that it is possible to speed up the optimization process by using lower-dimensional data representations at the beginning, while increasing the dimensionality of the input later in the optimization process. This is independent of the underlying optimization procedure, making the approach promising for many existing hyperparameter optimization algorithms.
74 citations
Cites background from "Neural Architecture Search with Rei..."
...Zoph et al.(31,32) train a recurrent neural network via reinforcement learning to ̄nd neural network architectures that are likely to yield a good performance on speci ̄c tasks....
[...]
...Zoph and Le (2016) train a recurrent neural network via reinforcement learning to find neural network architectures that are likely to yield a good performance on specific tasks....
[...]
References
More filters
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 2015TL;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 2014TL;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)....
[...]
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
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)....
[...]