Title:
TEXT CLASSIFICATION METHOD, APPARATUS, DEVICE, AND STORAGE MEDIUM
Document Type and Number:
WIPO Patent Application WO/2022/063202
Kind Code:
A1
Abstract:
A text classification method, an apparatus, a device, and a storage medium. Said method comprises: determining word vectors corresponding to various words in a text to be classified (S101); successively inputting the word vectors corresponding to the various words into a neuron model of a pulse neural network according to a text sequence, and successively processing the inputted various word vectors at various times (S102); and when the word vector currently processed by the neuron model is the last word vector in said text, determining a classification result of said text according to an output result of the neuron model (S103), wherein the neuron model determines an ith leakage coefficient according to an output result at the (i-1)th time and a target weight matrix, the ith leakage coefficient is used for determining an output result of the neuron model at the ith time, the value of the ith leakage coefficient is between 0 and 1, i is a sequence number of the processing time of the neuron model, and i is a positive integer greater than 1.
Inventors:
WU ZHENZHI (CN)
ZHU YAOLONG (CN)
ZHU YAOLONG (CN)
Application Number:
PCT/CN2021/120138
Publication Date:
March 31, 2022
Filing Date:
September 24, 2021
Export Citation:
Assignee:
LYNXI TECH CO LTD (CN)
International Classes:
G06F16/35
Foreign References:
| CN108984775A | 2018-12-11 | |||
| CN106875003A | 2017-06-20 | |||
| CN106875005A | 2017-06-20 | |||
| CN106875004A | 2017-06-20 | |||
| CN109754012A | 2019-05-14 | |||
| CN110826602A | 2020-02-21 | |||
| US20130204819A1 | 2013-08-08 |
Other References:
CHENG, LONG ET AL.: "Spiking Neural Networks: Model, Learning Algorithms and Applications", CONTROL AND DECISION, vol. 33, no. 5, 31 May 2018 (2018-05-31), pages 923 - 937, XP055914729, ISSN: 1001-0920
Attorney, Agent or Firm:
TEE&HOWE INTELLECTUAL PROPERTY ATTORNEYS (CN)
Download PDF: