Talk to books, play word association with Google AI
-
(11 days ago)
-
IANS
San Francisco, April 15 : Google has been working extensively on Artificial Intelligence (AI) and now it has given users a (fun) glimpse of how far natural language processing -- that deals with machine reading comprehension -- in the technology has come.
Google Research division of the search-giant has rolled out Semantic Experiences, which are websites with interesting activities that demonstrate AIs' ability to understand how we speak.
It has two experiences to enjoy and the third one is for developers to help them create their own experience.
The first two experiences are called "Talk to Books" in which users can explore a new way to interact with books, and "Semantris" where people can play word association games powered by semantic search.
In "Talk to Books" experience, users can simply type in a statement or a question and the AI will find whole sentences in books related to what they have typed.
Google Research Director of Engineering Ray Kurzweil and Product Manager Rachel Bernstein said the system does not depend on keyword matching.
They trained its AI by feeding it a "billion conversation-like pairs of sentences," so it can learn to identify what a good response looks like.
For example, if you type "Best detective in the world", the AI responds with several paragraphs and sentences that are related to the word "detective".
This way users can find exact lines from books which they mildly remember.
The second section Semantris offers word association games like a Tetris-like break-the-blocks 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 "Lionel Messi" 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.
According to Engadget, the development in word vector, an AI-training model that enables algorithms to learn relationships between words based on actual language usage, led to the advancement in natural language processing over the past few years.
Kurzweil and Bernstein said that these websites show how AIs' "new capabilities can drive applications that were not possible before".
They said other potential applications include "classification, semantic similarity, semantic clustering, whitelist applications (selecting the right response from many alternatives) and semantic search (of which Talk to Books is an example)."
Google CEO Sundar Pichai has been "betting big" on advances in AI and machine learning.
Earlier this year, Pichai said that AI is one of the most profound things that humanity is working on right now and compared it to basic utilities in terms of its importance.
"AI is 'one of the most important things that humanity is working on. It's more profound than, I don't know, electricity or fire," The Verge quoted Pichai as saying.
Pichai also said that AI could be used to help solve climate change issues or to cure cancer.
