Mark Zuckerberg creates AI, uses it to listen to Red Hot Chili Peppers
Mark Zuckerberg has been ploughing brain cells and money into building an artificial intelligence system for his home that can not only turn lights up or down but can also play Adele and Red Hot Chili Peppers on demand.
As pointed out by Ny Mag‘s Brian Feldman on Twitter, Zuckerberg posted a note to Facebook yesterday titled Building Jarvis, in which he details his personal challenge of 2016: “To build a simple AI to run my home – like Jarvis in Iron Man.”
In the post, Zuckerberg sets out the different artificial intelligence techniques used by the computer, including “natural language processing, speech recognition, face recognition, and reinforcement learning” and says that he has so far programmed Jarvis to control a number of different systems in the house.
However, in the Natural Language section of the post, Zuckerberg explains that music is a lot more complex than flicking a switch on and off, simply “because there are too many artists, songs and albums for a keyword system to handle. The range of things you can ask it is also much greater.” Source code commanding Jarvis to “play some Red Hot Chili Peppers” is then embedded in the post, with tracks including ‘Don’t Stop’ and ‘Under The Bridge’. Erm.
“When you say “play X”, even subtle variations can mean many different things,” he continues. “Consider these requests related to Adele: “Play someone like you”, “play someone like adele”, and “play some adele”. Those sound similar, but each is a completely different category of request. The first plays a specific song, the second recommends an artist, and the third creates a playlist of Adele’s best songs. Through a system of positive and negative feedback, an AI can learn these differences.”
He adds: “The more context an AI has, the better it can handle open-ended requests. At this point, I mostly just ask Jarvis to “play me some music” and by looking at my past listening patterns, it mostly nails something I’d want to hear.
“If it gets the mood wrong, I can just tell it, for example, “that’s not light, play something light”, and it can both learn the classification for that song and adjust immediately.”