Lightmatter, Inc.

The Charles Stark Draper Laboratory, Inc.

Tomo Lazovich

Darius Bunandar

Martin B Z Forsythe

Nicholas Christopher Harris

Tyler J Kenney

Michael Gould

Carl Ramey

Martin Forsythe

Ayon Basumallik

Onur Ozdemir

Rebecca Russell

David Walter

Louis Kim

Nicholas Dronen

Ludmila Levkova

Lakshmi Nair

Jacob Harer

David Widemann

Described herein are techniques of training a machine learning model and performing inference using an analog processor. Some embodiments mitigate the loss in performance of a machine learning model resulting from a lower precision of an analog processor by using an adaptive block floating-point representation of numbers for the analog processor. Some embodiments mitigate the loss in performance of a machine learning model due to noise that is present when using an analog processor. The techniques involve training the machine learning model such that it is robust to noise.

Machine learning model training using an analog processor

Systems and methods for performing signed matrix operations using a linear photonic processor are provided. The linear photonic processor is formed as an array of first amplitude modulators and second amplitude modulators, the first amplitude modulators configured to encode elements of a vector into first optical signals and the second amplitude modulators configured to encode a product between the vector elements and matrix elements into second optical signals. An apparatus may be used to implement a signed value of an output of the linear processor. The linear photonic processor may be configured to perform matrix-vector and/or matrix-matrix operations.

Systems and methods for analog computing using a linear photonic processor

Methods and apparatus for pre-processing first data for use with a trained machine learning model. In some embodiments, the method may comprise accessing the first data, wherein the first data has a first precision; generating, based on at least a first portion of the first data, second data having a second precision lower than the first precision; and providing the second data as input to the trained machine learning model to generate model output.

Quantized inputs for machine learning models

Techniques for computing matrix operations for arbitrarily large matrices on a finite-sized hybrid analog-digital matrix processor are described. Techniques for gain adjustment in a finite-sized hybrid analog-digital matrix processor are described which enable the system to obtain higher energy efficiencies, greater physical density and improved numerical accuracy. In some embodiments, these techniques enable maximization of the predictive accuracy of a GEMM-based convolutional neural network using low-precision data representations.

Hybrid analog-digital matrix processors

Aspects of the present application relate to techniques for computing convolutions and cross-correlations of input matrices. A first technique is based on the transformation of convolution operations into a matrix-vector product. A second technique is based on two-dimensional matrix multiplication. A third technique is based on the convolution theorem, which states that convolutions correspond to multiplications in a transform space. Embodiments include methods for computing convolutions of a filter matrix and an input data matrix; apparatuses for computing convolutions of a filter matrix and an input data matrix; and a non-transitory computer readable medium programmed with instructions that, when executed by a processor perform a method for computing convolutions of a filter matrix and an input data matrix.

Convolutional layers for neural networks using programmable nanophotonics

Systems and methods for performing matrix operations using a path-number balanced optical network are provided. The optical network is formed as an array including active optical components and passive optical components arranged at a substantially central location of the array. The optical network includes at least NM active optical components which are used to implement a first matrix of any size N&#xd7;M by embedding the first matrix in a second matrix of a larger size. The optical network performs matrix-vector and matrix-matrix operations by propagating one or more pluralities of optical signals corresponding to an input vector through the optical network.

Path-number-balanced universal photonic network

Methods and apparatus for training a matrix-based differentiable program using a photonics-based processor. The matrix-based differentiable program includes at least one matrix-valued variable associated with a matrix of values in a Euclidean vector space. The method comprises configuring components of the photonics-based processor to represent the matrix of values as an angular representation, processing, using the components of the photonics-based processor, training data to compute an error vector, determining in parallel, at least some gradients of parameters of the angular representation, wherein the determining is based on the error vector and a current input training vector, and updating the matrix of values by updating the angular representation based on the determined gradients.

Systems and methods for training matrix-based differentiable programs

A software instruction code repair system comprising an instruction code example pool. The example pool comprises a set of good instruction code examples and a set of bad instruction code examples. The software instruction code repair system further comprises a sequence-to-sequence (seq2seq) network that is configured to generate a corrected instruction code example, based on one example of the set of bad instruction code examples. The software instruction code repair system further comprises a discriminator configured to randomly select one of the corrected instruction code example and one instruction code example of the set of good instruction code examples to produce a selected instruction code example. The discriminator is further configured to make a determination that the selected instruction code example is most likely taken either the instruction code example pool or the seq2seq network.

Growing community of inventors

Cambridge, MA, United States of America

Tomo Lazovich