Contextual Summarisation Algorithm for summarising content in ANY language
Tuesday, April 12, 2016
We are four IIT Delhi graduates, who have developed the world's first contextual automatic summarisation algorithm which means that the summaries are not only depend on the article, but also on the past events that shape the information in the article. The algorithm gives 97% accuracy (information & flow), compared with 20,000 human generated summaries. We started building this because as students we were irritated by the amount of text we had to consume to understand the crux of a chapter. We tried a lot of different ways to solve this problem, we failed 3 times, but were able to create a game-changing algorithm over the course of 11 months. There are a lot of applications of this technology in verticals like - Education, Law, Research, Finance, Medicine, Daily Use (imagine opening multiple tabs after google search on mobile versus pressing and getting a summary and then deciding to go in), etc. We raised a seed round from HT Media & North Base Media (A US based, media focused, VC firm) and launched an summary news aggregation app (Zuppit : Live News Summary) recently. The central idea that drives us here is to save billions of man-hours by helping them wade off unnecessary content.