Related papers: All-Chalcogenide Programmable All-Optical Deep Neu…
Precise spatial and temporal manipulation of neural activity in specific genetically defined cell populations is now possible with the advent of optogenetics. The emerging field of optogenetics consists of a set of naturally-occurring and…
The recent explosive compute growth, mainly fueled by the boost of AI and DNNs, is currently instigating the demand for a novel computing paradigm that can overcome the insurmountable barriers imposed by conventional electronic computing…
Deep neural networks have become a pervasive tool in science and engineering. However, modern deep neural networks' growing energy requirements now increasingly limit their scaling and broader use. We propose a radical alternative for…
Deep neural networks, in particular convolutional neural networks, have become highly effective tools for compressing images and solving inverse problems including denoising, inpainting, and reconstruction from few and noisy measurements.…
All-optical active plasmonic devices are of fundamental importance for designing efficient nanophotonic circuits. We theoretically propose and numerically investigate an active plasmonic device made up of a nonlinear epsilon-near-zero…
Photonic computing using chalcogenide phase-change materials (PCMs) is under active development for energy-efficient artificial intelligence (AI) applications. A key requirement is to enable as many optically programmable levels per device…
This paper is the first attempt to use geometric deep learning and Sobolev training to incorporate non-Euclidean microstructural data such that anisotropic hyperelastic material machine learning models can be trained in the finite…
In visual recognition, the key to the performance improvement of ResNet is the success in establishing the stack of deep sequential convolutional layers using identical mapping by a shortcut connection. It results in multiple paths of data…
We propose all-optical neural networks characterized by very high energy efficiency and performance density of inference. We argue that the use of microcavity exciton-polaritons allows to take advantage of the properties of both photons and…
We find a heterogeneity in both complex and real valued neural networks with the insight from wave optics, claiming a much more important role of phase than its amplitude counterpart in the weight matrix. In complex-valued neural networks,…
In recent years, numerous research efforts have been dedicated towards developing efficient implantable devices for brain stimulation. However, there are limitations and challenges with the current technologies. Firstly, the stimulation of…
Current deep neural networks are highly overparameterized (up to billions of connection weights) and nonlinear. Yet they can fit data almost perfectly through variants of gradient descent algorithms and achieve unexpected levels of…
The explosive growth of data and its related energy consumption is pushing the need to develop energy-efficient brain-inspired schemes and materials for data processing and storage. Here, we demonstrate experimentally that Co/Pt films can…
The rapid growth of brain-inspired computing coupled with the inefficiencies in the CMOS implementations of neuromrphic systems has led to intense exploration of efficient hardware implementations of the functional units of the brain,…
A class of chalcogenide alloy materials that shows significant changes in optical properties upon an amorphous-to-crystalline phase transition has lead to development of large data capacities in modern optical data storage. Among…
The parallelism of optics and the miniaturization of optical components using nanophotonic structures, such as metasurfaces present a compelling alternative to electronic implementations of convolutional neural networks. The lack of a…
Optical neural networks (ONNs) perform extensive computations using photons instead of electrons, resulting in passively energy-efficient and low-latency computing. Among various ONNs, the diffractive optical neural networks (DONNs)…
All-optical switches have attracted attention because they can potentially overcome the speed limitation of electric switches. However, ultrafast, energy-efficient all-optical switches have been challenging to realize due to the…
While software implementations of neural networks have driven significant advances in computation, the von Neumann architecture imposes fundamental limitations on speed and energy efficiency. Neuromorphic networks, with structures inspired…
Photonic technologies offer great prospects for novel ultrafast, energy-efficient and hardware-friendly neuromorphic (brain-like) computing platforms. Moreover, neuromorphic photonic approaches based upon ubiquitous, technology-mature and…