3 October 2018 in Light: Science & Applications
Multimode fibers (MMFs) are an example of a highly scattering medium, which scramble the coherent light propagating within them to produce seemingly random patterns. Thus, for applications such as imaging and image projection through an MMF, careful measurements of the relationship between the inputs and outputs of the fiber are required. We show, as a proof of concept, that a deep neural network can learn the input-output relationship in a 0.75 m long MMF. Specifically, we demonstrate that a deep convolutional neural network (CNN) can learn the nonlinear relationships between the amplitude of the speckle pattern (phase information lost) obtained at the output of the fiber and the phase or the amplitude at the input of the fiber. Effectively, the network performs a nonlinear inversion task. We obtained image fidelities (correlations) as high as ~98% for reconstruction and ~94% for image projection in the MMF compared with the image recovered using the full knowledge of the system transmission characterized with the complex measured matrix. We further show that the network can be trained for transfer learning, i.e., it can transmit images through the MMF, which belongs to another class not used for training/testing.
15 December 2023 in Science
Recent successes in deep learning for vision and natural language processing are attributed to larger models but come with energy consumption and scalability issues. Current training of digital deep-learning models primarily relies on backpropagation that is unsuitable for physical implementation. In this work, we propose a simple deep neural network architecture augmented by a physical local learning (PhyLL) algorithm, which enables supervised and unsupervised training of deep physical neural networks without detailed knowledge of the nonlinear physical layer’s properties. We trained diverse wave-based physical neural networks in vowel and image classification experiments, showcasing the universality of our approach. Our method shows advantages over other hardware-aware training schemes by improving training speed, enhancing robustness, and reducing power consumption by eliminating the need for system modeling and thus decreasing digital computation.
3 September 2025 in Nature
15 July 2020 in Nature Machine Intelligence
3 September 2025 in Nature
Artificial intelligence (AI) and combinatorial optimization drive applications across science and industry, but their increasing energy demands challenge the sustainability of digital computing. Most unconventional computing systems1–7 target either AI or optimization workloads and rely on frequent, energy-intensive digital conversions, limiting efficiency. These systems also face application-hardware mismatches, whether handling memory-bottlenecked neural models, mapping real-world optimization problems or contending with inherent analog noise. Here we introduce an analog optical computer (AOC) that combines analog electronics and three-dimensional optics to accelerate AI inference and combinatorial optimization in a single platform. This dual-domain capability is enabled by a rapid fixed-point search, which avoids digital conversions and enhances noise robustness. With this fixed-point abstraction, the AOC implements emerging compute-bound neural models with recursive reasoning potential and realizes an advanced gradient-descent approach for expressive optimization. We demonstrate the benefits of co-designing the hardware and abstraction, echoing the co-evolution of digital accelerators and deep learning models, through four case studies: image classification, nonlinear regression, medical image reconstruction and financial transaction settlement. Built with scalable, consumer-grade technologies, the AOC paves a promising path for faster and sustainable computing. Its native support for iterative, compute-intensive models offers a scalable analog platform for fostering future innovation in AI and optimization.
11 September 2017 in IEEE Journal of Quantum Electronics
1-D compound metallic grating (CMG) is a periodic structure with more than one slit in each period. When CMG is combined with a graphene sheet as its cover, the incident light is effectively coupled to the plasmons in graphene which in turn can result in strong manipulation of light for both major polarizations. We show that tunable phase resonance and perfect absorption of the incident light are interesting outcomes of this manipulation. In this paper, we demonstrate that fano-like phase resonances which can be observed in CMGs under transverse magnetic polarized incident wave are tuned by changing the Fermi level of graphene. It is shown that while the spectral position of the phase resonances can be shifted up to several gigahertzes, their peak to peak amplitudes are tuned from ~0.9 to ~0.1. On the other hand, we design a graphene-covered CMG, which is able to perfectly absorb both major polarizations of incident wave in two separate bands; hence, providing the opportunity for designing multi-band/wide-band absorbers. We have developed a circuit model for the analysis of the structure. Parameters of the model are derived explicitly and analytically for both major polarizations. Our results are verified through comparison against results of the full-wave simulations.