OpenAI's New Bot Defeats Some of the World's Best Dota 2 Players

August 16, 2017

With recent victories over the world's best Go players and inroads into the realm of StarCraft 2 making headlines, it should come as no surprise that artificial intelligence (A.I.) is one of the hottest topics in the shared space between gaming and the development of machine learning and other cutting edge processes. However, these milestones could pale in comparison to the impressive work of Elon Musk's OpenAI team at Valve's annual Dota 2 tournament.

So why exactly did the OpenAI team make a visit to one of the most exciting e-sports events of the year? As The Verge's T.C. Sottek reports, this nonprofit joined the festivities for a surprise segment that pitted it’s A.I. against some of the best players in the global Dota 2 competitive scene. The end result? An astounding display of skill that left leading names, like Danylo "Dendi" Ishutin, both in awe and summarily defeated.

In terms of how this bot was able to topple some of the most accomplished human players and potentially raise the bar for A.I. performance in gaming, Sottek goes on to explain that it all boils down to a mere two weeks of real-time learning. Of course, this "two weeks" of training is a measurement from a human perspective, and the members of the OpenAI research unit have pointed out that this period actually equated to "lifetimes" of learning and experience for the neural network behind the company's Dota 2 bot.

While most might see this exhibition as nothing more than a tech-oriented parlor trick designed to excite the online gaming community, the truth of the matter is that testing the boundaries of A.I. and developing critical thinking skills via neural networks could open the door to other real-life applications and uses. From self-driving cars to the technology that could eventually replace most humans in the workforce, advancements in A.I. such as this one prove time and time again that machine learning and neural networks hold the key to the future of countless industries.

Could this program be the foundation on which researchers develop the next great breakthrough in A.I. systems? How does this test fit into Elon Musk's warnings regarding the dangers of unregulated A.I. development? For answers to these questions, and many more, feel free to follow along with Sottek for the full story on the other side of the link below.

Read the Full Article Here