Title:
DECISION MODEL TRAINING METHOD AND APPARATUS, COMPUTER DEVICE, AND STORAGE MEDIUM
Document Type and Number:
WIPO Patent Application WO/2023/138155
Kind Code:
A1
Abstract:
A decision model training method and apparatus, a computer device, and a storage medium, relating to the field of artificial intelligence. The method comprises: acquiring a first model pool, wherein the first model pool comprises at least one decision model, and the decision model is used for indicating a battle strategy used by a virtual character during a battle (310); in an n-th round of iteration process, inputting i-th battle state data into an n-th decision model to obtain an action instruction output by the n-th decision model, wherein the action instruction is used for instructing a first client to control the first virtual character to execute an action, and to feed back (i+1)-th battle state data collected after the first virtual character executes the action (320); training the n-th decision model on the basis of battle state data in a battle process to obtain a (n+1)-th decision model (330); and adding the (n+1)-th decision model to the first model pool (340). By training a decision model by means of battle state data in a battle process, the battle anthropomorphism of a decision model is improved.
Inventors:
LIU RUOCHEN (CN)
CAO QIYANG (CN)
LIU SZE YEUNG (CN)
ZHANG LIANGPENG (CN)
ZENG ZHENGWEN (CN)
CAO QIYANG (CN)
LIU SZE YEUNG (CN)
ZHANG LIANGPENG (CN)
ZENG ZHENGWEN (CN)
Application Number:
PCT/CN2022/129092
Publication Date:
July 27, 2023
Filing Date:
November 01, 2022
Export Citation:
Assignee:
TENCENT TECH SHENZHEN CO LTD (CN)
International Classes:
A63F13/55
Foreign References:
| CN114404976A | 2022-04-29 | |||
| CN112870721A | 2021-06-01 | |||
| CN111330279A | 2020-06-26 | |||
| CN112295237A | 2021-02-02 | |||
| CN113827946A | 2021-12-24 | |||
| US20190111344A1 | 2019-04-18 |
Attorney, Agent or Firm:
BEIJING SAN GAO YONG XIN INTELLECTUAL PROPERTY AGENCY CO., LTD. (CN)
Download PDF: