| Autor: |
Huang, Yanmei, Mei, Qiang, Hu, Mulan, Vadivel, Thanjai, Raj, A. Daison |
| Předmět: |
|
| Zdroj: |
International Journal of Speech Technology; Jun2022, Vol. 25 Issue 2, p421-433, 13p |
| Abstrakt: |
The Video Game has been seen as the most demanding classical intelligence game for a long time search and difficulties in determining node location and movement. A new approach, Voice Assisted Virtual Game Architecture, is presented in this research to decide node assignments using Strategy Networks and Valuation Network (VVGA-SNVN) to pick movements. In this framework, Deep Neural networks are formed through a new combination of supervised learning and human sports reinforcement for studying self-play games. The neural networks play on the most advanced tree search programs simulating dozens of arbitrary self-play games without looking away. Besides, a new search algorithm combines the simulation of the algorithmic tree with strategy and valuation networks. This search's algorithm is thus obtained a 93% victory in terms of execution accuracy over other virtual games. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
Full text from SpringerLink