Experimentally realized in situ backpropagation for deep learning in photonic neural networks.

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 ([Formula: see text]94% test accuracy), and energy scaling analysis indicated a route to scalable machine learning.

Experimental realization of arbitrary activation functions for optical neural networks.

We experimentally demonstrate an on-chip electro-optic circuit for realizing arbitrary nonlinear activation functions for optical neural networks (ONNs). The circuit operates by converting a small portion of the input optical signal into an electrical signal and modulating the intensity of the remaining optical signal. Electrical signal processing allows the activation function circuit to realize any optical-to-optical nonlinearity that does not require amplification. Such line shapes are not constrained to those of conventional optical nonlinearities. Through numerical simulations, we demonstrate that the activation function improves the performance of an ONN on the MNIST image classification task. Moreover, the activation circuit allows for the realization of nonlinearities with far lower optical signal attenuation, paving the way for much deeper ONNs.

Use of sentinel serology in a Bovine viral diarrhea virus eradication program.

Forty-seven cattle management groups from 36 herds in a regional Bovine viral diarrhea virus (BVDV) eradication program were selected to evaluate serology as a tool to detect herd infection with BVDV. Serum samples were obtained from 5 non-vaccinated sentinel calves ≥ 6 months old in each management group and virus neutralizing (VN) antibody titers against BVDV genotypes 1 and 2 were determined. A herd was considered positive if 2 or more sentinel calves had VN antibody titers ≥ 128 to either genotype. Results were compared to individual animal testing of all available calves by reverse transcription polymerase chain reaction (RT-PCR) on skin biopsy samples. In 1 management group from 1 herd (n = 24), 3 sentinel calves had VN antibody titers ≥ 128. Three ear notch samples from that herd were positive for BVDV on RT-PCR assay. All other management groups were negative for BVDV. In the present study, the herd sensitivity of sentinel serology was 100% (95% confidence interval [CI]: 0.05-1.0) and herd specificity was 100% (95% CI: 0.90-1.0). The κ value for agreement between sentinel serology and RT-PCR was 1.0 (95% CI: 1.0-1.0). Preliminary results suggest that sentinel animal serology can be utilized in a BVDV eradication program to provide an accurate and efficient evaluation of herd status.