Paper Reading #23: Aspect-level News Browsing: Understanding News Events from Multiple Viewpoints (IUI 38)

Comments

http://stuartjchi.blogspot.com/2011/04/paper-reading-21-addressing-problems-of.html?

http://shennessy11.blogspot.com/2011/04/paper-reading-21.html

Reference Information

   Title:Aspect-level News Browsing: Understanding News Events from Multiple Viewpoints

   Author:Souneil Park, SangJeong Lee, Junehwa Song

   Publisher: IUI '10, February 7-10, 2010 Hong Kong

Summary

Bias in news is a major obstacle in keeping the public well informed. The reality of the situation is is that all journalists that write for news outlets carry some sort of bias, where it be in the framing of the situation, the aspects covered, the tone in which they cover them and the style in which they write. Journalists are continually influenced by political and ideological views as well as external factors of news production such as owners and advertisers. It is difficult for readers to build an entire balanced understanding of a news event without seeking many different sources.

The researchers in this article aim to provide a computational solution to the media bias problem. They developed a framework that mitigates the effects of media bias. It creates and provides readers with multiple classified viewpoints on news events. The core function of this framework is "aspect-level news browsing." Aspect-level news browsing gives readers a classified view of articles with many different viewpoints and facilitates interactions that help them to discover and compare existing biased views of an event.

It allows them to understand events from a plural of viewpoints in order to formulate their own balanced opinion. They first demonstrated the feasibility of this idea through user studies. They then developed a framing cycle-aware cluster. The results showed that the method performs classification more accurately than other methods.

Discussion

I read the news a lot and something like this would be nice to have. That said, I don't like reading a lot of opinionated articles on both sides of the fence. I wonder if this system grabs the extreme sides rather than something a little more neutral. I feel like those articles that don't consider anything else get annoying and just go back and forth. This system would be nice if implemented correctly, but I think i'll stick with Google News.