TRAINING DEVICE, INFERENCE DEVICE, TRAINING METHOD, INFERENCE METHOD, PROGRAM, AND COMPUTER-READABLE NON-TRANSITORY STORAGE MEDIUM

Title:

TRAINING DEVICE, INFERENCE DEVICE, TRAINING METHOD, INFERENCE METHOD, PROGRAM, AND COMPUTER-READABLE NON-TRANSITORY STORAGE MEDIUM

Document Type and Number:

WIPO Patent Application WO/2021/025075

Kind Code:

A1

Abstract:

[Problem] To achieve a high-precision simulator through machine learning. [Solution] An inference device is provided with one or more memories and one or more processors. The one or more processors are configured so as to input a time-series input variable and a latent variable which is the output of a first network to the first network and extract a latent variable for the current step, and to input the latent variable to a second network and infer a time-series state with respect to the input variable.

Inventors:

YOSHIKAWA MASASHI (JP)
SAKAI MASAHIRO (JP)

Application Number:

PCT/JP2020/030065

Publication Date:

February 11, 2021

Filing Date:

August 05, 2020

Export Citation:

Click for automatic bibliography generation Help

Assignee:

PREFERRED NETWORKS INC (JP)

International Classes:

G06N3/04

Foreign References:

US20190050734A1	2019-02-14
US20190108436A1	2019-04-11

Other References:

ZHANG DINGSHENG, RUBANOVA YULIA, BETTENCOURT JESSE, DUVENAUD DAVID: "Neural ordinary differential equations", ARXIV.ORG/ABS/1806.07366V1, 19 June 2018 (2018-06-19), XP055791967, Retrieved from the Internet [retrieved on 20201026]
YOTA MIZUTANI; YOSHIMASA TSURUOKA: "Introducing an Environment Model based on Hidden States for Model-Based Deep Reinforcement Learning", GAME PROGRAMMING WORKSHOP 2018 PROCEEDINGS; NOVEMBER 16-18, 2018, 9 November 2018 (2018-11-09), pages 72 - 79, XP009526765

Attorney, Agent or Firm:

NAKAMURA Yukitaka et al. (JP)

Download PDF:

View/Download PDF PDF Help

Previous Patent: GROUP CALLING SYSTEM, GROUP CALLING METHOD, AND PROGRAM

Next Patent: PULSED HIGH FREQUENCY MONITOR