More and more news consumers are feeling news fatigue. Navigating all of the sources can become so complicated that some consumers that they just give up trying. Francis Tseng has secured a Knight Foundation Prototype Grant in order to fix that. It’s called Argos.
The basic idea of Argos is this:
- Identify stories, topics and entities within them. Keep track of those aspects of stories in some kind of database.
- Use automation to track updates to those stories and isolate the parts that are new, based on (for now) six to 10 of the most well known news sources.
- Create bare bones briefings for new arrivals to a story that don’t require clicking link after link to find out the basic idea of who the entities are.
- Deliver all that to users through a web and iOS interface.
Many news stories are parts of larger stories. So, for example, if you are following the debacle around HealthCare.gov, you are also following a part of the Obamacare story. Tseng described the problem of creating an algorithm that can decipher the links between those stories as a software problem, but helping users of Argos understand the relationship is really a design problem. Tseng works in both fields.
Tseng told Technically Brooklyn that his goal for the prototype process is to get a basic version of the app up and on a handful of people’s devices so he can start testing and refining it. “It is going quicker than I anticipated,” he said, “which is good.”
How could machine learning digest text-based news and sort out which bits are updates and which rehash previous stories? This is not wholly new ground for Tseng. As a designer and developer, he’s been playing around with these ideas for a while.
The structure of news helps him, he said. A good rule of thumb for assessing how well an algorithm is at summarizing a piece of text is to compare it to the baseline of the piece’s first two sentences. Since so much news puts its most important information in those first two lines, it is tough to beat that baseline. “You can almost get away with just snatching that first paragraph,” Tseng said.
Tseng started seriously working on Argos around August 2013. He lives and works from Williamsburg.
He is working on a limited basis for his employer IDEO now, as he spends almost all of his time on developing Argos as a team, for now, of one.
His vision for the app is big, though he has no previous background in news. He told Technically Brooklyn, “From what I’ve read, journalism tends to sort of have this strong moral foundation to it, in a sense that it’s a very important institution in a democracy. Keeping people informed is its duty. … But people have grown to neglect that a bit or not appreciate it as much, so now the industry is sort of flailing, or withering a bit.”
“I’m hoping that Argos, the application, can restore some of that former glory to the profession or to the practice.-30-