Intel Corporation

Daniel David Ben-Dayan Rubin

Ehud Cohen

Michael Behar

Dmitri Vainbrand

Yonatan Glesner

Elad Hoffer

There is disclosed in an example, a processor, having: decode circuitry to decode instructions from an instruction stream; a data cache unit including circuitry to cache data for the processor; and a compute unit having an approximate matrix multiplication (AMM) circuit comprising: a data receptor to receive a weight vector w and an input vector x, both of size N, and a compression regulating parameter n; and a factorizor circuit to factorize w into w&#x2245;B&#xb7;s, by computing a binary factorized matrix B of size N&#xd7;n, and a dictionary vector s of size n. In an example, the factorization follows a dual minimization procedure, the time complexity of which is on average linear with N.

Binary vector factorization

A processor, including: decode circuitry to decode instructions; a data cache unit including circuitry to cache data for the processor; and an approximate matrix multiplication (AMM) circuit including: a data receptor circuit to receive a weight vector w and an input vector x, both of size N, and a compression regulating parameter n; a factorizer circuit to factorize w into w&#x2245;B&#xb7;s, by computing a binary factorized matrix B of size N&#xd7;n, and a dictionary vector s of size n; and a binary multiplier circuit to compute w^T x&#x2245;(B&#xb7;s)^T x=s^T(B^T x), the binary multiplier circuit comprising a hardware accelerator circuit to compute an array productB^T x).

Binary multiplier for binary vector factorization

Techniques are provided for image segmentation based on image differencing, using recursive neural networks. A methodology implementing the techniques according to an embodiment includes quantizing pixels of a first image frame, performing a rigid translation of the quantized first image frame to generate a second image frame, and performing a differencing operation between the quantized first image frame and the second image frame to generate a sparse image frame. A neural network can then be applied to the sparse image frame to generate a segmented image. In still another embodiment, the methodology is applied to a sequence or set of image frames, for example from a video or still camera, and pixels from a first and second image frame of the sequence/set are quantized. The sparse image frame is generated from a difference between quantized image frames. The method further includes training the neural network on sparse training image frames.

Image difference based segmentation using recursive neural networks

A processor includes a linear approximator and a front end including circuitry to assign linear approximation of a nonlinear function to a linear approximator. The linear approximator includes circuitry to divide a range of values for the linear approximation into a defined number of segments, perform linear approximation for each segment, move borders between the segments to reduce discontinuity moving along segments of variable length, repeat linear approximation for each segment until convergence, and return values for the linear approximation.

Growing community of inventors

Tel Aviv, Israel

Daniel David Ben-Dayan Rubin