Invention Grant
- Patent Title: Compressed recurrent neural network models
-
Application No.: US15172457Application Date: 2016-06-03
-
Publication No.: US10515307B2Publication Date: 2019-12-24
- Inventor: Tara N. Sainath , Vikas Sindhwani
- Applicant: Google LLC
- Applicant Address: US CA Mountain View
- Assignee: Google LLC
- Current Assignee: Google LLC
- Current Assignee Address: US CA Mountain View
- Agency: Fish & Richardson P.C.
- Main IPC: G06F15/18
- IPC: G06F15/18 ; G06N3/08 ; G06N3/04

Abstract:
Methods, systems, and apparatus, including computer programs encoded on computer storage media, for implementing long-short term memory layers with compressed gating functions. One of the systems includes a first long short-term memory (LSTM) layer, wherein the first LSTM layer is configured to, for each of the plurality of time steps, generate a new layer state and a new layer output by applying a plurality of gates to a current layer input, a current layer state, and a current layer output, each of the plurality of gates being configured to, for each of the plurality of time steps, generate a respective intermediate gate output vector by multiplying a gate input vector and a gate parameter matrix. The gate parameter matrix for at least one of the plurality of gates is a structured matrix or is defined by a compressed parameter matrix and a projection matrix.
Public/Granted literature
- US20170076196A1 COMPRESSED RECURRENT NEURAL NETWORK MODELS Public/Granted day:2017-03-16
Information query