Deep learning with coherent nanophotonic circuits
01 Jul 2017-Vol. 11, Iss: 7, pp 441-446
TL;DR: A new architecture for a fully optical neural network is demonstrated that enables a computational speed enhancement of at least two orders of magnitude and three order of magnitude in power efficiency over state-of-the-art electronics.
Abstract: Artificial Neural Networks have dramatically improved performance for many machine learning tasks. We demonstrate a new architecture for a fully optical neural network that enables a computational speed enhancement of at least two orders of magnitude and three orders of magnitude in power efficiency over state-of-the-art electronics.
Citations
More filters
TL;DR: In this paper, the concept of an operator parameter space consisting of physical entities encoded with Maxwell's equations is introduced to improve the network's capability to generalize beyond their training set, and the network trained with operator parameters yields remarkably accurate predictions of the topological transitions both within and beyond the training physical space.
Abstract: Neural networks based on machine learning can interpolate well within the training dataset, but their ability to extrapolate is severely limited by fundamental issues such as the bias-variance trade-off. Here we introduce the concept of an operator parameter space consisting of physical entities encoded with Maxwell's equations to improve the networks' capability to generalize beyond their training set. We illustrate the idea with photonic crystals, and show that the network trained with operator parameters yields remarkably accurate predictions of the topological transitions both within and beyond the training physical space. Such concepts can be generalized to higher-dimensional wave structures by choosing the appropriate operator parameters.
20 citations
TL;DR: In this article, a vaccination strategy was proposed to improve the robustness of diffractive networks against 3D misalignments and fabrication tolerances in the physical implementation of a trained diffractive network.
Abstract: As an optical machine learning framework, Diffractive Deep Neural Networks (D2NN) take advantage of data-driven training methods used in deep learning to devise light-matter interaction in 3D for performing a desired statistical inference task. Multi-layer optical object recognition platforms designed with this diffractive framework have been shown to generalize to unseen image data achieving e.g., >98% blind inference accuracy for hand-written digit classification. The multi-layer structure of diffractive networks offers significant advantages in terms of their diffraction efficiency, inference capability and optical signal contrast. However, the use of multiple diffractive layers also brings practical challenges for the fabrication and alignment of these diffractive systems for accurate optical inference. Here, we introduce and experimentally demonstrate a new training scheme that significantly increases the robustness of diffractive networks against 3D misalignments and fabrication tolerances in the physical implementation of a trained diffractive network. By modeling the undesired layer-to-layer misalignments in 3D as continuous random variables in the optical forward model, diffractive networks are trained to maintain their inference accuracy over a large range of misalignments; we term this diffractive network design as vaccinated D2NN (v-D2NN). We further extend this vaccination strategy to the training of diffractive networks that use differential detectors at the output plane as well as to jointly-trained hybrid (optical-electronic) networks to reveal that all of these diffractive designs improve their resilience to misalignments by taking into account possible 3D fabrication variations and displacements during their training phase.
20 citations
TL;DR: In this article, a photonic extreme learning machine (PEM) is proposed for machine learning using optical encoder and coherent wave propagation in free space, with the far field acting as feature mapping space.
Abstract: Photonic brain-inspired platforms are emerging as novel analog computing devices, enabling fast and energy-efficient operations for machine learning. These artificial neural networks generally require tailored optical elements, such as integrated photonic circuits, engineered diffractive layers, nanophotonic materials, or time-delay schemes, which are challenging to train or stabilize. Here we present a neuromorphic photonic scheme - photonic extreme learning machines - that can be implemented simply by using an optical encoder and coherent wave propagation in free space. We realize the concept through spatial light modulation of a laser beam, with the far field that acts as feature mapping space. We experimentally demonstrated learning from data on various classification and regression tasks, achieving accuracies comparable to digital extreme learning machines. Our findings point out an optical machine learning device that is easy-to-train, energetically efficient, scalable and fabrication-constraint free. The scheme can be generalized to a plethora of photonic systems, opening the route to real-time neuromorphic processing of optical data.
20 citations
TL;DR: This article reviews the existing research regarding the use of ML in nano-scale biomedical engineering and identifies the main challenges that can be formulated as ML problems, and discusses the state of the art ML methodologies used to countermeasure the aforementioned challenges.
Abstract: Machine learning (ML) empowers biomedical systems with the capability to optimize their performance through modeling of the available data extremely well, without using strong assumptions about the modeled system. Especially in nano-scale biosystems, where the generated data sets are too vast and complex to mentally parse without computational assist, ML is instrumental in analyzing and extracting new insights, accelerating material and structure discoveries and designing experience as well as supporting nano-scale communications and networks. However, despite these efforts, the use of ML in nano-scale biomedical engineering remains still under-explored in certain areas and research challenges are still open in fields such as structure and material design and simulations, communications and signal processing, and bio-medicine applications. In this article, we review the existing research regarding the use of ML in nano-scale biomedical engineering. In more detail, we first identify and discuss the main challenges that can be formulated as ML problems. These challenges are classified in three main categories: structure and material design and simulation, communications and signal processing and biomedicine applications. Next, we discuss the state of the art ML methodologies that are used to countermeasure the aforementioned challenges. For each of the presented methodologies, special emphasis is given to its principles, applications and limitations. Finally, we conclude the article with insightful discussions, that reveal research gaps and highlight possible future research directions.
20 citations
Cites background from "Deep learning with coherent nanopho..."
...Nanophotonic systems are also utilized for neuromorphic computing and especially for the realization of deep learning networks [109] and adsorptionbased photonic NNs [110]....
[...]
10 May 2019
TL;DR: A protocol for training photonic neural networks based on adjoint methods by physically backpropagating an optical error signal and calculating the gradient of the network with respect to its tunable degrees of freedom.
Abstract: We provide a protocol for training photonic neural networks based on adjoint methods. The gradient of the network with respect to its tunable degrees of freedom is computed by physically backpropagating an optical error signal. © 2019 The Author(s)
20 citations
Cites methods from "Deep learning with coherent nanopho..."
...Although we focus our discussion on one particular recently proposed hardware implementation [3], our conclusions are derived starting from Maxwell’s equations, and the ideas could therefore extend to other photonic neural network platforms, as well as to other applications....
[...]
...[3], we assume that the matrix-vector multiplications are implemented using an Optical Interference Unit (OIU), consisting of a mesh of reconfigurable Mach-Zehnder interferometers....
[...]
References
More filters
Proceedings Article•
03 Dec 2012TL;DR: The state-of-the-art performance of CNNs was achieved by Deep Convolutional Neural Networks (DCNNs) as discussed by the authors, which consists of five convolutional layers, some of which are followed by max-pooling layers, and three fully-connected layers with a final 1000-way softmax.
Abstract: We trained a large, deep convolutional neural network to classify the 1.2 million high-resolution images in the ImageNet LSVRC-2010 contest into the 1000 different classes. On the test data, we achieved top-1 and top-5 error rates of 37.5% and 17.0% which is considerably better than the previous state-of-the-art. The neural network, which has 60 million parameters and 650,000 neurons, consists of five convolutional layers, some of which are followed by max-pooling layers, and three fully-connected layers with a final 1000-way softmax. To make training faster, we used non-saturating neurons and a very efficient GPU implementation of the convolution operation. To reduce overriding in the fully-connected layers we employed a recently-developed regularization method called "dropout" that proved to be very effective. We also entered a variant of this model in the ILSVRC-2012 competition and achieved a winning top-5 test error rate of 15.3%, compared to 26.2% achieved by the second-best entry.
73,978 citations
[...]
TL;DR: Deep learning is making major advances in solving problems that have resisted the best attempts of the artificial intelligence community for many years, and will have many more successes in the near future because it requires very little engineering by hand and can easily take advantage of increases in the amount of available computation and data.
Abstract: Deep learning allows computational models that are composed of multiple processing layers to learn representations of data with multiple levels of abstraction. These methods have dramatically improved the state-of-the-art in speech recognition, visual object recognition, object detection and many other domains such as drug discovery and genomics. Deep learning discovers intricate structure in large data sets by using the backpropagation algorithm to indicate how a machine should change its internal parameters that are used to compute the representation in each layer from the representation in the previous layer. Deep convolutional nets have brought about breakthroughs in processing images, video, speech and audio, whereas recurrent nets have shone light on sequential data such as text and speech.
46,982 citations
TL;DR: This work bridges the divide between high-dimensional sensory inputs and actions, resulting in the first artificial agent that is capable of learning to excel at a diverse array of challenging tasks.
Abstract: The theory of reinforcement learning provides a normative account, deeply rooted in psychological and neuroscientific perspectives on animal behaviour, of how agents may optimize their control of an environment. To use reinforcement learning successfully in situations approaching real-world complexity, however, agents are confronted with a difficult task: they must derive efficient representations of the environment from high-dimensional sensory inputs, and use these to generalize past experience to new situations. Remarkably, humans and other animals seem to solve this problem through a harmonious combination of reinforcement learning and hierarchical sensory processing systems, the former evidenced by a wealth of neural data revealing notable parallels between the phasic signals emitted by dopaminergic neurons and temporal difference reinforcement learning algorithms. While reinforcement learning agents have achieved some successes in a variety of domains, their applicability has previously been limited to domains in which useful features can be handcrafted, or to domains with fully observed, low-dimensional state spaces. Here we use recent advances in training deep neural networks to develop a novel artificial agent, termed a deep Q-network, that can learn successful policies directly from high-dimensional sensory inputs using end-to-end reinforcement learning. We tested this agent on the challenging domain of classic Atari 2600 games. We demonstrate that the deep Q-network agent, receiving only the pixels and the game score as inputs, was able to surpass the performance of all previous algorithms and achieve a level comparable to that of a professional human games tester across a set of 49 games, using the same algorithm, network architecture and hyperparameters. This work bridges the divide between high-dimensional sensory inputs and actions, resulting in the first artificial agent that is capable of learning to excel at a diverse array of challenging tasks.
23,074 citations
"Deep learning with coherent nanopho..." refers background or methods in this paper
...The computational resolution of ONNs is limited by practical non-idealities, including (1) thermal crosstalk between phase shifters in interferometers, (2) optical coupling drift, (3) the finite precision with which an optical phase can be set (16 bits in our case), (4) photodetection noise and (5) finite photodetection dynamic range (30 dB in our case)....
[...]
...(3) Once a neural network is trained, the architecture can be passive, and computation on the optical signals will be performed without additional energy input....
[...]
...We used four instances of the OIU to realize the following matrix transformations in the spatial-mode basis: (1) U((1))Σ((1)), (2) V((1)), (3) U((2))Σ((2)) and (4) V((2))....
[...]
...Transformations (1) and (2) realize the first matrix M((1)), and (3) and (4) implement M((2))....
[...]
TL;DR: In this article, an effective way of initializing the weights that allows deep autoencoder networks to learn low-dimensional codes that work much better than principal components analysis as a tool to reduce the dimensionality of data is described.
Abstract: High-dimensional data can be converted to low-dimensional codes by training a multilayer neural network with a small central layer to reconstruct high-dimensional input vectors. Gradient descent can be used for fine-tuning the weights in such "autoencoder" networks, but this works well only if the initial weights are close to a good solution. We describe an effective way of initializing the weights that allows deep autoencoder networks to learn low-dimensional codes that work much better than principal components analysis as a tool to reduce the dimensionality of data.
16,717 citations
TL;DR: This historical survey compactly summarizes relevant work, much of it from the previous millennium, review deep supervised learning, unsupervised learning, reinforcement learning & evolutionary computation, and indirect search for short programs encoding deep and large networks.
14,635 citations
"Deep learning with coherent nanopho..." refers methods in this paper
...ANNs can be trained by feeding training data into the input layer and then computing the output by forward propagation; weighting parameters in each matrix are subsequently optimized using back propagation [16]....
[...]