Friday, June 10, 2011

The Semantic Barrier!

Mapping Social Media (watch the video here)

Tom Thai from Bluefin Labs explains how they are connecting the semantic gap.
Shout out to Tom Thai for using a hockey reference in his speech too, even though he is a Bruins fan #GoCanucks 

OK, this technology is pretty crazy, and not even in the sense that it is extremely complicated to understand for humans, but because it is making it possible for computers to make human inferences and connections between social media conversations and how they are/were sparked. They have a sweet data slice visualization of 50,000 people. relevant vocab, social graph: which is made up of all your Facebook friends, Twitter followers etc. Content graph: everything that happens on TV. Episodes of TV, a commercial that runs inside of a TV show would be connected to that TV show.

The guys at Bluefin Labs have come up with this technology that focuses on connecting the feedback loop between audience and speaker but on a grand scale using new technology. Or in other words, they have put in 15 years of research at MIT which is allowing them to map out social media chatter and see where it stems from, what started it, what it is relative to. Here is a good quote from the article that explains what the challenge is:

"the challenge in all of this: help machines understand the semantic layer, or the what the conversation is about. As he describes it, the first input is TV, the second input is social media. The space in the middle is the semantic barrier."

They call this mapping between the social and content graph the TV Genome, which is cool? maybe giving themselves a little too much credit with that name haha. Anyways, it does seem to have some futuristic implications. The article even gets a little profane with one of its examples so dive in if you life the F-word. If not, I would still suggest watching the video! 
The take home message from the article is that by harvesting this data on such a HUGE level, and doing it correctly with computers, will allow us to discover new surprising insights from the data. gotta love insights. I don't know about you, but if you are in advertising and you went to Washington State you probably enjoy reading Mintel in your free time ;)

Thoughts? what could the implications be? what could be discovered? is this even worth doing?


No comments:

Post a Comment