Despite Google dominating the Online search market, every now & then there are some who venture out in this field & launch new search engines. Some succeed, some obviously sink into oblivion. So, once again, there’s a new search engine called ThisPlusThat.me. Not only is the name different, the search engine, too, works on a different methodology.
ThisPlusThat.me is the brainchild of a Ph.D. holding data scientist Christopher Moody while he was part of the Insights Data Science Fellow Program, a six week post doctorate training fellowship.
Christopher blogs on the Insights website: Natural language “isn’t that great for searching. When you type a search query into Google, you miss out on a wide spectrum of human concepts and human emotions. ThisPlusThat thus captures the relationships between words. This goes beyond sentiment analysis, or parts-of-speech tagging — words, products, and locations are learned from context.
The new search engine’s filters are based on a principle quite different than those of traditional search engines. The latter work such that although they delete most of the fluff, they do not understand the relationships between words, and understand only the similarities or dissimilarities.
Christopher’s search engine works on this principle. He initially started experimenting with the word2vec algorithm by Google, which allows the addition & deletion of concepts as if they were vectors, & get out “sensible and interesting results.” word2vec is a type of distributed word representation algorithm that trains a neural network in order to assign a vector to every word.
Here’s an example given by Christopher: word2vec allows users to evaluate a query like King – Man + Woman = Queen. Or, one can type in Bill Gates without Microsoft and see what results come up.
Interesting is the first adjective that comes to mind when one starts using this new search engine. The main question of course is – will it stand up there, shoulder to shoulder, with Google?
Image Credit: ThisPlusThat.me
Advertising Message
