eWEEK content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.
Microsoft acquired Mojang, the Swedish game company behind Minecraft, the popular block-based sandbox construction game, for a whopping $2.5 billion in 2014. As it turns out, Microsoft not only purchased the game and inherited the loyal community that sprouted around it, the software giant also ended up with a virtual laboratory in which to test artificial intelligence (AI) software.
This week, the company peeled the curtain back on Project AIX. Hailing from Microsoft Research, the platform uses the Minecraft to conduct AI research. Currently, AIX is only available to select academic researchers participating in a private beta of the software. Microsoft plans to release AIX this summer under an open-source license.
More than just a game, Minecraft is evolving into a gateway to other computing experiences.
During last year’s E3 trade show in Los Angeles, Microsoft showed off an early build of Minecraft running on HoloLens, the company’s augmented reality headset. In January, the company snapped up MinecraftEdu from Teacher Gaming LLC to help bring the education-themed build of Minecraft to more classrooms.
While Minecraft’s blocky, Lego-like world seems deceptively simple on the surface, intrepid gamers have cobbled together complex systems, including working computers and virtual hard drives, with an astute understanding of the game’s rules and clever placement of in-game materials. Microsoft is hoping that some of that ingenuity rubs off on future AI technologies.
AIX is a ‘mod’, gamer speak for a modification that alters the content of a video game, for the Java version of Minecraft and along with code that enables artificial intelligence agents to sense and act within the game’s environment. The software runs on Windows, Mac OS X or Linux. Users can program their agents in any programming language.
Microsoft researchers have been using AIX to train a Minecraft character to scale a hill practically unaided. They are programming it essentially to learn instead of “programming it to accomplish specific tasks,” Fernando Diaz, a senior researcher in the Microsoft Research New York City lab, said in a blog post detailing Project AIX.
Though it may stem from a hit game, AIX is not a consumer product, Microsoft said. Instead, the company is hoping the technology will attract academic researchers and motivated amateurs.
Microsoft isn’t alone in its pursuit of smarter, more self-sufficient AI.
In December, Facebook announced it had open-sourced its GPU-based Big Sure AI servers after having worked on the technology for 18 months. The hardware uses graphical processing units (GPUs) from Nvidia’s Tesla Accelerated Processing Platform to speed up the training of neural networks, the building blocks of AI technologies.
Also last year, Dell announced a major, multibillion dollar investment in the Chinese technology market that includes partnering with the Chinese Academy of Science to create a joint lab tasked with growing work on AI and advanced computing technologies. In February, IBM unveiled new image recognition and emotional IQ APIs that help sharpen Watson’s visual and emotional senses.