Artificial Intelligence and Video Games Development

NB: We do not resell your papers. Upon ordering, we do an original paper exclusively for you.

NB: All your data is kept safe from the public.

Artificial Intelligence: Article Review

In their article, Vinyals, Gaffney, and Ewalds (2017) discuss a collaboration in which Blizzard Entertainment, a video game developer, and DeepMind, an artificial intelligence (AI) company, have been using video games as a platform for AI design and testing. The research environment is based on StarCraft II, a highly popular real-time strategy (RTS) video game (“StarCraft AI competition,” 2018).

DeepMind’s researchers worked with Blizzard to design a machine learning API that would allow the AI developers to use the game to test their agents. The education of these agents took the form of missions where the AI acted as players, collecting materials, completing campaigns, and playing mini-games (DeepMind, 2017). The mini-games, however, have served mainly as a basis for further exploration, and players as well as researchers have been encouraged to create new games and missions for the agents to complete.

This opportunity to make use of games with a heavy focus on strategy and long-term planning has allowed AI developers to test their current knowledge and abilities in an area requiring multiple agents to complete various interrelated tasks. According to Peng et al. (2017), such environments as StarCraft are better suited for complex testing than games that have been previously employed such as Atari.

Here, the concept of collaboration is leading to an increased quality of communication and scalability for agents (Peng et al., 2017). This joint effort between the companies is of interest as it introduces the gamification of research in an environment with a large number of human players (Blizzard Entertainment, 2017). Therefore, in contrast to settings that have been designed for agents only, StarCraft and Blizzard can offer DeepMind an enormous amount of data gathered from playing time which teaches the AI to perform a set of complex tasks.

From an AI perspective, the concept of utilizing games for development and testing of machine learning holds tremendous potential which made this article inherently interesting. The complexity of overlaying game elements such as map awareness, unit health, and strategic movement can be compared to the intricacy of real-world scenarios. Application of AI mechanisms in various industries ranging from economics to military requires analysis of big data in real-time with consequent decision-making. The utilization of the DeepMind in combination with SC2LE tools holds promising results in AI research such as sequence prediction and long-term memory which can expand possibilities in AI applications for highly complex tasks.