Why did The Huffington Post buy Adaptive Semantics? It comes down to competitive advantage.
The online publisher indicated that other licensees of its JuLiA technology (that’s JuLiA on the left, by the way), such as CNN and Disqus, won’t likely have their contracts renewed, and you can put that down to the fact that Huff Post is heavily invested in continually boosting its strong online community advantage. “Community is extremely important to The Huffington Post,” says Adaptive Semantics co-founder Jeff Revesz, who informs us that The Huffington Post is actually up to 2.8 million comments a month â€“ “more than any other publisher out there and it’s looking at growing that more in the future, in growing the social and content graph. That needs serious machine learning work in the background powering applications on top of that,” he says.
As director of social news, Revesz expects to have even more opportunity to stay focused on machine learning R&D to drive those apps than when he was trying to run a business, too. On the comment moderation front, Adaptive Semantics has gone from creating a series of ad hoc models and algorithms that were informed by stricter understandings of inappropriate content to a more customized solution that let publishers better manage to their own editorial standards, submitting their own data and training their own algorithms to that end, he says. But the other step â€“ and an even bigger one for how The Huff Post can leverage its JuLiA platform and its advances in machine learning algorithms â€“ has been Adaptive Semantics’ progress on applications around expert discovery.
Engage and Reward
“We started to think about how to highlight content, engage with editors who want to build applications for citizen journalism, and how we are able to pick out the best contributions to community and reward users,” Revesz says. That’s building up the social graph. About six months ago it began working at being able to use its technology to help analyze the data to determine who is submitting highest-quality content, and now its machine learning capabilities are important to the publisher’s newly introduced badge system for more deeply engaging readers and giving them something in return for their community contributions (such as comment moderation privileges). Other editorial systems and the stats they deliver around replies, fans, friends and the like play a role in badges, as well, as Revesz admits that no publisher would be comfortable having machine learning completely drive such efforts. “But it brings tremendous efficiency to the process,” he says.
This latest acquisition in the semantics space points to how the hype around semantics and machine learning is turning into real viable solutions, he thinks. “From a market standpoint there’s room for other machine learning companies that have focus beyond public relations and analyzing sentiment of tweets to show the world what they think of a new razor,” he says. “I really feel personally that this acquisition shows there is room in the market for general machine learning shops, if you have a creative and interesting application.”
At all of age 27, Revesz is thrilled to have sold his first company â€“ and we think that qualifies him, along with similarly co-founder and COO Elena Haliczer, to be added to our new crop of young guns driving the semantic web.
As we asked here, please let us know about other young professionals you know about who are having a big impact in the semantic web â€“ we’d love to bring their stories to light.
â€¢ And now that you’ve seen the success of this year and a half old startup, don’t forget to propose your startup for our Semantic Web Impact Awards. The deadline is Sept. 15.