Yes, you heard it right! Search giant Google’s new AI experiment has you talking to books. Sounds kind of absurd, right? Well, this is proof that Google’s work in natural language understanding is getting better all the time!
Google has been working on Artificial Intelligence for quite some time, and now it has given the world a small fun glimpse of how far they’ve reached. Google’s Research Division announced three new AI experiments under “Semantic Experiences” that lets web users fiddle with semantics and natural language processing.
One of the experiences is called “Talk to Books”, where users can explore a new fun way to interact with books. The second experience is called “Semantris” where people can play fun word association games. The last of the three experiences is for developers, where Google has shared some of the models for natural language processing, which the developers can use! You can see them here.
In “Talk to Books”, you can literally converse with a machine learning-training algorithm, that provides answers to questions with relevant passages from human-written text. Damn, if that is not cool, I don’t know what is! In layman ‘s terms, each time you make a statement or ask a question, the tool finds sentences in books that respond, with no dependence on the matching keyword. So, to know if I’m interested in reading the book or not, I can just ask the book and get a relevant response!
The second experience “Semantris” is a place where people can play word association games powered by semantic search. This includes a Tetris like break-the-blocks game. In this game experience, the AI would display random blocks with text written on them. Users have to “break” those blocks by typing a word which can relate to any of the text written on them. For example, if the AI displays “Football” on any block, users could write “Cristiano Ronaldo” in the space provided below the blocks. The AI processes and matches the word with the text displayed on the blocks. Once the AI matches with the word, it “breaks” the block and some points are awarded.
For all the geeks out there, the primary sub-field of the AI Google is developing is called as “word vectors”. It’s a natural language understanding which maps semantically similar phrases to nearby points based on equivalence, similarity or relatedness of ideas and language. As Google’s director of engineering Ray Kurzweil states in his blogpost, “it’s a way to enable algorithms to learn about the relationships between words, based on examples of actual language usage”.
Google has also published its work on the topic in a paper here. It’s also made a pre-trained module available on its TensorFlow platform for researchers and developers to dabble with.
Well, I can’t wait how far this technology reaches the end users! I’m so glad to be in what is called as the “Google” era!