Game On! – Tiago Pinheiro Teixeira, Programmer, Developer, and Producer in the Gaming Industry – Game Development, Deep Learning and AI: What’s Here and What’s Coming Soon in the Gaming Industry and How Will It Impact Other Technology Devel

Tiago Pinheiro Teixeira, a programmer, developer, and producer in the gaming industry, delivers a thorough overview of technological advances in the gaming industry that engage deep learning. He describes the development process and testing phase of games and how games can even be used to teach and train artificial intelligence (AI).

Gaming strategy and development often utilizes traditional decision tree processing. Teixeira discusses how some gaming companies are now training their agents using machine learning models, researching the use of AI to develop better dialogue within the games for more realistic interactions, and of course utilizing AI for testing and production. Regarding dialogue, Teixeira explains that much more work is needed in the area, for as virtual assistant bots need only answer a user’s question, game characters must also role-play the character with accents and motivations, which is far more advanced than simple Q & A. As Teixeira states, currently, AI is used for functional decision-making such as where to go, how to hide, how to shoot, etc. Additionally, he discusses how games are now being used to train real world AI, for example, training fully autonomous vehicles to function properly in their required tasks. By presenting scenarios via games, real world AI can improve its responses to unpredictable situations that could occur, which could help make products and experiences involving AI more sophisticated and safer.

Teixeira talks about other gaming experiments that have shown and produced interesting results. He discusses a particular experiment at Google Brain, a deep learning artificial intelligence research team at Google that managed to teach AI to work in teams, which is an incredibly complex issue in deep learning. And the possibilities from this new development are immense, as this knowledge can potentially be used in a myriad of areas with specialized AI now working together to accomplish complex tasks or solve problems in general. Teixeira expects to see incredible developments in the coming decade.

The gaming expert explains how games that require long-term, or real time strategy, beyond the simple, move here, fire there, type actions, may use deep learning to train from scratch, to accomplish more involved tasks. And Teixeira mentions that there is an abundance of API (application programming interface) online for the general public to access, download, and conduct their own experimentation. API essentially assists developers in the building of applications. By exposing only the specific objects or actions a developer needs, API can dramatically simplify programming.

Testing is the key to success. For as Tiago Pinheiro Teixeira describes, gaming developers may have a team of testers working daily, playing the games and finding bugs for six months or more, but this pales in comparison to millions of gamers playing all around the globe. True balance of a character statistically comes more often after launch, after millions of gamers delve into the game deeply.