IBM vs. Twilight fans: ‘Breaking Dawn’ breaks emotional analytics

The problem with Twilight is, apparently, that it's too emotionally complex - Or, at least, it is if you're hoping to use audience reactions as a way of teaching computers how to process emotions.

The final installment of the Twilight Saga series of movies, The Twilight Saga: Breaking Dawn Part 2, didn’t just rule the box office over the Thanksgiving weekend, it also ruled social media, with more than 5 million Tweets devoted to its mix of horror, soap opera and teenage angst. However, the content of those Tweets wasn’t necessarily as straightforward as you might expect, and that complexity caused problems for IBM when it tried to analyze the online chatter.

IBM researchers worked with analysts from USC Annenberg to look at the Twitter commentary on Breaking Dawn Part 2 and other movies in the two weeks leading up to the Thanksgiving weekend with the intent of using the messages to “teach” its computer systems to be able to more easily recognize and process emotional responses. The problem was, Twilight fans were a little bit more emotional than IBM techs had anticipated, and expressed those emotions in a way that made them less easily “read” by automated systems.

The difficulty was first noticed when researchers realized that Twilight was ranking noticeably lower than other movies in terms of the “social sentiment” surrounding the movie, with fans apparently less happy about the movie following its release than they were beforehand (Pre-release, Breaking Dawn had 90 percent positive Tweets, with that figure dropping to 75 percent following its debut in theaters; by contrast, Life of Pi scored 91 percent positive Tweets, and Rise of The Guardians an impressive 99 percent positive). It wasn’t that fans were unhappy with the quality of the movie, it turns out, but that they were sad about the movie series ending with this installment.

According to the write-up in the Hollywood Reporter, applying “analystics and natural language processing technologies to study the Tweets” revealed that “some of the sentiment was reflecting the ‘audiences’ emotional reaction to the tear-jerking moments in the movie.’” The website goes on to talk about the analysis with IBM’s general manager for Media and Entertainment, Steve Canepa, who says that automating the collation and accurate analysis of audience reaction through social media to an entertainment event is going to become increasingly important to studios and other content creators as Twitter, Facebook and other social outlets allow audiences to share their reactions faster than old-fashioned polling methods can measure. “What we are doing is essentially creating a focus group in real time,” Canepa explained, adding “We are attempting to understand what are they saying, what motivates people to see [a] movie.”

Of course, currently it looks as if this would be much easier if only Hollywood would make movies that elicited only one emotional reaction in audiences, as opposed to multiple emotions that could possibly conflict with each other.

View the original article here