scispace - formally typeset
Search or ask a question
Journal ArticleDOI

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
Journal ArticleDOI
TL;DR: Photonic integrated circuits (PIC) have recently attracted extensive attention in advanced photonic signal processing to meet the ever-increasing demands on high-speed and ultra-compact data signal processing.
Abstract: Photonic integrated circuits (PICs) have recently attracted extensive attention in advanced photonic signal processing to meet the ever-increasing demands on high-speed and ultracompact data signal...

14 citations

Journal ArticleDOI
TL;DR: In this paper, a Si microring resonator (MRR) crossbar arrays are used as a programmable nanophotonoic processor (PNP) for a deep learning accelerator.
Abstract: We propose Si microring resonator (MRR) crossbar arrays as a programmable nanophotonoic processor (PNP) for a deep learning accelerator. The proposed MRR crossbar array can perform multiply–accumulate (MAC) operation in an optical domain. We numerically reveal that an optical neural network (ONN) based on the proposed MRR crossbar arrays can be used for the pattern recognition task with a similar performance to that of an ONN based on cascaded Mach–Zehnder interferometers (MZIs). We predict that the power consumption can be reduced by approximately tenfold. The small area of the MRR also contributes to the reduction in its chip size by a factor of 36. The fabricated test devices consisting of Si MRRs with phase shifters exhibited no significant crosstalk between neighboring MRRs and showed the feasibility of MAC operation. Photonic integrated circuit (PIC) using the proposed MRR crossbar arrays is promising for large-scale and low-power PNP for deep learning.

13 citations

Journal ArticleDOI
TL;DR: An electrodynamic model of the interacting, momentum-dependent polarizability for a two-dimensional metal on a model metallic substrate is derived, which quantitatively captures the evolution of the dielectric properties of borophene as a function of metal-borophene distance.
Abstract: We compute the dielectric properties of freestanding and metal-supported borophene from first-principles time-dependent density functional theory. We find that both the low- and high-energy plasmons of borophene are fully quenched by the presence of a metallic substrate at borophene-metal distances smaller than ≃9 A. Based on these findings, we derive an electrodynamic model of the interacting, momentum-dependent polarizability for a two-dimensional metal on a model metallic substrate, which quantitatively captures the evolution of the dielectric properties of borophene as a function of metal-borophene distance. Applying this model to a series of metallic substrates, we show that maximizing the plasmon energy detuning between borophene and substrate is the key material descriptor for plasmonic performance.

13 citations

Proceedings ArticleDOI
05 Dec 2021
TL;DR: CrossLight as discussed by the authors proposes a cross-layer optimized neural network accelerator called CrossLight that leverages silicon photonics, including device-level engineering for resilience to process variations and thermal crosstalk, circuit-level tuning enhancements for inference latency reduction, and architecture-level optimizations to enable better resolution, energy-efficiency, and throughput.
Abstract: Domain-specific neural network accelerators have seen growing interest in recent years due to their improved energy efficiency and performance compared to CPUs and GPUs. In this paper, we propose a novel cross-layer optimized neural network accelerator called CrossLight that leverages silicon photonics. CrossLight includes device-level engineering for resilience to process variations and thermal crosstalk, circuit-level tuning enhancements for inference latency reduction, and architecture-level optimizations to enable better resolution, energy-efficiency, and throughput. On average, CrossLight offers 9.5x lower energy-per-bit and 15.9x higher performance-per-watt than state-of-the-art photonic deep learning accelerators.

13 citations

Journal ArticleDOI
17 May 2022-Science
TL;DR: In situ backpropagation is demonstrated for the first time to solve classification tasks and a new protocol to keep the entire gradient measurement and update of physical device voltages in the analog domain is evaluated, improving on past theoretical proposals.
Abstract: Integrated photonic neural networks provide a promising platform for energy-efficient, high-throughput machine learning with extensive scientific and commercial applications. Photonic neural networks efficiently transform optically encoded inputs using Mach-Zehnder interferometer mesh networks interleaved with nonlinearities. We experimentally trained a three-layer, four-port silicon photonic neural network with programmable phase shifters and optical power monitoring to solve classification tasks using “in situ backpropagation,” a photonic analog of the most popular method to train conventional neural networks. We measured backpropagated gradients for phase-shifter voltages by interfering forward- and backward-propagating light and simulated in situ backpropagation for 64-port photonic neural networks trained on MNIST image recognition given errors. All experiments performed comparably to digital simulations (>94% test accuracy), and energy scaling analysis indicated a route to scalable machine learning. Description Editor’s summary Commercial applications of machine learning (ML) are associated with exponentially increasing energy costs, requiring the development of energy-efficient analog alternatives. Many conventional ML methods use digital backpropagation for neural network training, which is a computationally expensive task. Pai et al. designed a photonic neural network chip to allow efficient and feasible in situ backpropagation training by monitoring optical power passing either forward or backward through each waveguide segment of the chip (see the Perspective by Roques-Carmes). The presented proof-of-principle experimental realization of on-chip backpropagation training demonstrates one of the ways that ML could fundamentally change in the future, with most of the computation taking place optically. —Yury Suleymanov Nanophotonic neural networks can be trained using on-chip backpropagation, unlocking opportunities for analog computing.

13 citations

References
More filters
Proceedings Article
03 Dec 2012
TL;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

Journal ArticleDOI
28 May 2015-Nature
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

Journal ArticleDOI
26 Feb 2015-Nature
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))....

    [...]

Journal ArticleDOI
28 Jul 2006-Science
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

Journal ArticleDOI
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]....

    [...]