From the outside in

Tuesday, February 8, 2011

Semantic Text Startup Reads and Summarizes Documents for You

via Mashable! by Jennifer Van Grove on 2/7/11


The Spark of Genius Series highlights a unique feature of startups and is made possible by Microsoft BizSpark. If you would like to have your startup considered for inclusion, please see the details here.

Name: Topicmarks

Quick Pitch: Topicmarks summarizes your cloud documents and extracts the most important points for you.

Genius Idea: If you’ve ever come across a document too large to digest in a few minutes, you’re likely to find a new friend in Topicmarks.

The startup’s beta product represents more than four years of machine learning and artificial intelligence research that can statistically analyze cloud documents, or those you upload to service, to quickly distill them down to six sections: Overview, Facts, Summary, Keywords, Index and Properties.

The technology is so adept at reading text and understanding context that Topicmarks can cut up the full text within documents into sentences, assign meaning to words and pluck out relevant facts, all within minutes. Think of it as a better, machine-enabled version of CliffsNotes — but for all text documents.

Topicmarks users can upload documents, use the bookmarklet to grab web content, e-mail in text, or copy and paste text for nearly instant analysis. Key points are extracted and personalized to the user.

The company is also working on an Evernote plugin capable of automatically tagging new notes, applying tags to old notes and adding-in quick links to condensed summaries. Currently, users can already import their Evernote notes to view summaries.

Topicmarks technology could also be applied to distilling key points from social media streams or RSS feeds, avenues that the startup may choose to pursue at a later date. For now, the immediate focus is integrating with popular cloud services such as Dropbox or Scribd and improving the design of the site to make it more appealing to end users.

The product, which has an infinite number of use cases, is primarily targeted at the more than 50 million information workers in the world producing or reproducing information, says CEO Roland Siebelink.

Interestingly enough, the technology is already proving popular with students and teachers, though a few in the latter group have e-mailed to complain that their students are “cheating” with the help of Topicmarks. Siebelink tells the story of a disgruntled teacher who awarded one sneaky student with the highest mark on a class assignment, only to later discover that the summary had been copied word for word from Topicmarks.

This real-life example proves the quality of Topicmarks’ results, says Siebelink, who believes the tool ultimately helps both students and education workers.

Topicmarks is gaining interest from both end users and investors after winning two startup competitions — the winning Founder Showcase pitch is embedded below. The startup is reviewing term sheets from institutional investors and is said to be going after $1 million to $2 million in Series A funding, according to a source familiar with the matter.

The product is free for the time being, but the company will enact a freemium pricing structure in the future that may charge to process larger documents, says Siebelink.

Image courtesy of iStockphoto, thesuperph.


Series Supported by Microsoft BizSpark

Microsoft BizSpark

The Spark of Genius Series highlights a unique feature of startups and is made possible by Microsoft BizSpark, a startup program that gives you three-year access to the latest Microsoft development tools, as well as connecting you to a nationwide network of investors and incubators. There are no upfront costs, so if your business is privately owned, less than three years old, and generates less than U.S.$1 million in annual revenue, you can sign up today.

More About: evernote, founder showcase, machine learning, semantic analysis, spark-of-genius, startup, topicmarks

For more Startups coverage:

Posted via email from The New Word Order

No comments:

Post a Comment