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
27 Oct 2020
TL;DR: A computational approach for the large-scale detection of protein binding sites, that considers protein conformations as 3D-images, binding sites as objects on these images to detect, and conformational ensembles of proteins as3D-videos to analyze.
Abstract: Identification of novel protein binding sites expands druggable genome and opens new opportunities for drug discovery. Generally, presence or absence of a binding site depends on the three-dimensional conformation of a protein, making binding site identification resemble the object detection problem in computer vision. Here we introduce a computational approach for the large-scale detection of protein binding sites, that considers protein conformations as 3D-images, binding sites as objects on these images to detect, and conformational ensembles of proteins as 3D-videos to analyze. BiteNet is suitable for spatiotemporal detection of hard-to-spot allosteric binding sites, as we showed for conformation-specific binding site of the epidermal growth factor receptor, oligomer-specific binding site of the ion channel, and binding site in G protein-coupled receptor. BiteNet outperforms state-of-the-art methods both in terms of accuracy and speed, taking about 1.5 minutes to analyze 1000 conformations of a protein with ~2000 atoms.
32 citations
20 May 2019
TL;DR: The proposed RoemNet is a novel meta learning based channel estimation approach that involves a meta-learner that can learn from the environment of different channels and outperforms existing channel estimation methods including existing learning methods under various scenarios.
Abstract: Recently, in order to achieve performance improvement in scenarios where the channel is either unknown, or too complex for an analytical description, Neural Network (NN) based channel estimation is introduced in Orthogonal Frequency Division Multiplexing (OFDM) systems However, this kind of learning method is not reliable enough when the conditions of online deployment of the common NNs are not consistent with the channel models used in the training stage Furthermore, common NNs need plenty of data as well as time to be trained, which is not suitable for the OFDM communication network with time varying channels To tackle these challenges, we propose a novel meta learning based channel estimation approach called RoemNet The most distinctive characteristic of RoemNet is that it involves a meta-learner that can learn from the environment of different channels With the update of meta-learner, RoemNet is robust enough to solve new channel learning tasks using only a small number of pilots Furthermore, RoemNet can alleviate the effect of Doppler spread and significantly improve the Bit Error Ratio (BER) performance under different channel environments Experiment results demonstrate that the proposed RoemNet outperforms existing channel estimation methods including existing learning methods under various scenarios
32 citations
Cites background from "Neural Architecture Search with Rei..."
..., various basic parameters during neural network training, such as initialization of parameters [17], the choice of optimizers [18], even the structure of models [19], etc....
[...]
15 Jul 2019
TL;DR: The current progress in the neural architecture search and the implementation on hardware is reviewed and a multi-objective optimization problem, which aims to identify the best network and hardware pair to maximize accuracy with guaranteed hardware efficiency is discussed.
Abstract: Neural Architecture Search (NAS), that automatically identifies the best network architecture, is a promising technique to respond to the ever-growing demand for application-specific Artificial Intelligence (AI). On the other hand, a large number of research efforts have been put on implementing and optimizing AI applications on the hardware. Out of all leading computation platforms, Field Programmable Gate Arrays (FPGAs) stand out due to its flexibility and versatility over ASICs and its efficiency over CPUs and GPUs. To identify the best neural architecture and hardware implementation pair, a number of research works are emerging to involve the awareness of hardware efficiency in the NAS process, which is called "hardware-aware NAS". Unlike the conventional NAS with a mono-criteria of accuracy, hardware-aware NAS is a multi-objective optimization problem, which aims to identify the best network and hardware pair to maximize accuracy with guaranteed hardware efficiency. Most recently, the co-design of neural architecture and hardware has been put forward to further push forward the Pareto frontier between accuracy and efficiency trade-off. This paper will review and discuss the current progress in the neural architecture search and the implementation on hardware.
32 citations
Cites background from "Neural Architecture Search with Rei..."
...In [10], each cell is a normal convolution operation....
[...]
...Figure 3 shows the NAS framework presented in [10]....
[...]
...Existing work has demonstrated that the automatically generated network architectures can achieve close accuracy to the best human-invented architectures on the image classification task [10, 12]....
[...]
...For [10], it searches 20,000 networks and trains these networks across 500 P100 GPUs over 4 days....
[...]
...One of the most successful work is carried out by Google [10], which leverages reinforcement learning to predict hyperparameters to identify specific neural architectures, and utilizes their accuracies as rewards....
[...]
Posted Content•
[...]
TL;DR: It is argued that equal attention, if not more, should be paid to teaching, and furthermore, an optimization framework should be used to obtain good teaching strategies to achieve teacher-student co-evolution.
Abstract: Teaching plays a very important role in our society, by spreading human knowledge and educating our next generations. A good teacher will select appropriate teaching materials, impact suitable methodologies, and set up targeted examinations, according to the learning behaviors of the students. In the field of artificial intelligence, however, one has not fully explored the role of teaching, and pays most attention to machine \emph{learning}. In this paper, we argue that equal attention, if not more, should be paid to teaching, and furthermore, an optimization framework (instead of heuristics) should be used to obtain good teaching strategies. We call this approach `learning to teach'. In the approach, two intelligent agents interact with each other: a student model (which corresponds to the learner in traditional machine learning algorithms), and a teacher model (which determines the appropriate data, loss function, and hypothesis space to facilitate the training of the student model). The teacher model leverages the feedback from the student model to optimize its own teaching strategies by means of reinforcement learning, so as to achieve teacher-student co-evolution. To demonstrate the practical value of our proposed approach, we take the training of deep neural networks (DNN) as an example, and show that by using the learning to teach techniques, we are able to use much less training data and fewer iterations to achieve almost the same accuracy for different kinds of DNN models (e.g., multi-layer perceptron, convolutional neural networks and recurrent neural networks) under various machine learning tasks (e.g., image classification and text understanding).
32 citations
TL;DR: A survey of data augmentation for digital images in deep learning is presented in this paper, where the authors discuss the current image data augmentation techniques which can be used to increase the number of images.
Abstract: Deep learning proved its efficiency in many fields of computer science such as computer vision, image classifications, object detection, image segmentation, and more. Deep learning models primarily depend on the availability of huge datasets. Without the existence of many images in datasets, different deep learning models will not be able to learn and produce accurate models. Unfortunately, several fields don't have access to large amounts of evidence, such as medical image processing. For example. The world is suffering from the lack of COVID-19 virus datasets, and there is no benchmark dataset from the beginning of 2020. This pandemic was the main motivation of this survey to deliver and discuss the current image data augmentation techniques which can be used to increase the number of images. In this paper, a survey of data augmentation for digital images in deep learning will be presented. The study begins and with the introduction section, which reflects the importance of data augmentation in general. The classical image data augmentation taxonomy and photometric transformation will be presented in the second section. The third section will illustrate the deep learning image data augmentation. Finally, the fourth section will survey the state of the art of using image data augmentation techniques in the different deep learning research and application.
32 citations
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)....
[...]