Google Inc.

Cell Design Labs, Inc.

University of California

Anselm Levskaya

Lukasz Mieczyslaw Kaiser

Peter Emtage

Nikita Kitaev

Amy E Gilbert

Vladimir Anatolievich Slepushkin

Spencer Scott

Jonathon Shlens

Wendell A Lim

Christopher A Voigt

Niki J Parmar

Ashish Teku Vaswani

Irwan Bello

Prajit Ramachandran

A system implemented as computer programs on one or more computers in one or more locations that implements a computer vision model is described. The computer vision model includes a positional local self-attention layer that is configured to receive an input feature map and to generate an output feature map. For each input element in the input feature map, the positional local self-attention layer generates a respective output element for the output feature map by generating a memory block including neighboring input elements around the input element, generates a query vector using the input element and a query weight matrix, for each neighboring element in the memory block, performs positional local self-attention operations to generate a temporary output element, and generates the respective output element by summing temporary output elements of the neighboring elements in the memory block.

Fully attentional computer vision

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for performing a machine learning task on a network input to generate a network output. In one aspect, one of the systems includes an attention neural network configured to perform the machine learning task, the attention neural network including one or more LSH attention layers, each LSH attention layer comprising one or more LSH attention sub-layers, each LSH sub-layer configured to: receive a sequence of queries derived from an input sequence to the LSH attention layer, the sequence of queries having a respective query at each of a plurality of input positions; determine one or more respective hash values for each of the respective queries at each of the plurality of input positions; generate a plurality of LSH groupings; and generate an attended input sequence.

Attention neural networks with locality-sensitive hashing

The present invention relates to methods and compositions for reducing the immunogenicity of chimeric Notch receptors, and specifically to transcription factors useful for controlling gene expression delivered to tissues by such chimeric Notch receptors.

Methods and compositions for reducing the immunogenicity of chimeric notch receptors

The invention provides methods, materials and systems of regulating association between proteins of interest using light. In an aspect, the invention takes advantage of the ability of phytochromes to change conformation upon exposure to appropriate light conditions, and to bind in a conformation-dependent manner to cognate proteins called phytochrome-interacting factors. The invention comprises a method of regulating interaction between a first protein of interest and second protein within a cell by light. Such a method optionally comprises providing in the cell (1) a first protein construct which comprises the first protein, a phytochrome domain (PHD), and (2) providing in the cell a second protein construct which comprises the second protein and a phytochrome domain-interacting peptide (PIP) that can bind selectively to the Pfr state, but not to the Pr state, of the phytochrome domain.

Spatio-temporal control of protein interactions using phytochromes

Growing community of inventors

Oakland, CA, United States of America

Anselm Levskaya