Shulou(Shulou.com)11/24 Report--

CTOnews.com, December 10, prior to the announcement of Tencent's "Enlightenment" platform, it announced that it would open the application of "Arena of Valor AI Open Research Environment" for machine learning algorithm research for non-commercial uses. The "Enlightenment" platform will provide a unique high-complexity MOBA training environment unique to the industry, helping to facilitate cutting-edge exploration in related academic fields.

The latest open application, "Arena of Valor AI Open Research Environment", includes Arena of Valor's 1v1 combat environment and baseline algorithm model. Currently, it supports mirror combat tasks and non-mirror combat tasks of 20 heroes, which is especially suitable for studying generalization and generality in competitive reinforcement learning.

Now, researchers only need to register an account on the official website of Tencent Enlightenment platform, submit materials and pass the platform review, then they can download Arena of Valor 1v1 test environment gamecore for free and conduct AI-related research and testing locally.

With the growth and iteration of Arena of Valor games in continuous operation, the "Enlightenment" platform will gradually explore and open more game test environments suitable for AI research, providing a new solution to the problem of "lack of high-challenge test environment" faced by the AI industry.

In 2019, Arena of Valor and Tencent AI Lab jointly created an open research platform for "Enlightenment" AI. "Enlightenment" has officially opened Arena of Valor's 1v1 environment, and encapsulated the environment for ease of use, providing a standard code and training framework, so that individual users can easily get started with distributed training when resources permit.

CTOnews.com has learned that the latest open application, "Arena of Valor AI Open Research Environment", only considers the choice of heroes on both sides, supports 400 combat sub-tasks, and if the summoner skills are included, there will be 40000 seed tasks, which is a great challenge to the generalization of the model, so it is especially suitable for studying generalization and generality in competitive reinforcement learning.

Furthermore, the generalization of the model to different tasks is very important to reduce the application cost of AI in different fields. For example, in the face of multiple tasks in the test environment, how to get the best results under the limited resources and time conditions, the research methodology and results can be transferred to other fields, such as the control of different parts of the robot.

"Enlightenment" is an open research platform for AI multi-agent and complex decision-making led by Tencent. Relying on the core advantages of Tencent AI Lab and "Arena of Valor" in algorithms, computing power and experimental scenarios, it is an open research and application exploration platform for academic researchers and algorithm developers.

Based on the "enlightenment" platform, Tencent will join hands with academia and industry to make use of the game environment to continuously improve AI capabilities, and to find the possibility of AI technology to solve more problems, and play a greater role in robot, medical, industrial, transportation and other fields.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.