As a forerunner in artificial intelligence (AI) learning, Google extends its lead through the introduction of new ‘Semantic Experiences’. The company is announcing Semantic Experiences, a website that uses algorithms to understand the relationship between words. The website consists of two experiences offerings: Talk to Books and Semantris.

Talk to Books is a different take on how to explore books. There is no need to look at book covers or sift through pages. Ask a question and the tool responds with sentences from books as answers. For example, ask, “What is the meaning of life?” and the system responds with a quote from the book, Meanings of Life. Here’s the excerpt:

“A meaning of life can be considered as the outcome of a negotiation between the individual and the social system.”

Talk to Books learns from a billion conversation-like sentences, unlike Google’s traditional keyword-based platform. In that sense, one has to essentially ‘talk’ with books to find one that is interesting. The tool chooses from over 100,000 books, all of which are available as a response to any statement.

Despite the exciting prospects, Google stresses that the tool is not perfect. Talk to Books processes sentences that will not fit the context of full paragraphs. However, the tool remains a great example on how natural language processing is progressing; on many levels, it also confirms that search tools are still evolving.

Anyone keen on trying out Talk to Books can just click here.

Next up is Semantris – a word association game that also utilises the same technology. The game comes in two modes, Arcade and Blocks.

In Arcade, the system will arrange words vertically, split by a blue line. The player must find a word that has high association with those at the top. Once entered, the system rearranges the words by association. Scoring is then provided for words that fall under the blue line.

The second mode, Blocks, will be more familiar to those who have tried Talk to Books. The game will arrange blocks in a Tetris-style format. Players will use either words or phrases to destroy word-associated blocks. For example, to destroy a ‘horse’ block, one can write ‘a four-legged creature’. For the Blocks mode, Google encourages that users try to be as obscure as possible in writing their hints.

Keen on  trying out Semantris, then click here.

(Source: Google)