interesting use of crowdsourcing for storytelling

[Kurt Luther]

This seems closely related to our project:


http://www.darpa.mil/NewsEvents/Releases/2013/05/14.aspx

Riedl used the hypothetical scenario of a bank robbery as a test case
for collecting stories and generating a plot graph. In the example, a
would-be bank robber named John (1) drives to the bank, (2) enters the
bank, and then, (3) faced with a plot decision, either sees Sally (the
bank teller), waits in line or scans the bank. At such decision
points, the narrative can split based on which actions the
contributors agreed would follow as a result of the player’s choice.
John and Sally’s interactions unfold through a series of such decision
points until, as all of the narrative lines agree, John ultimately
leaves the bank, at which point he is either arrested or gets away,
and the story concludes.

Because some plot events cannot logically co-occur in a single
narrative, the system identifies mutual exclusion between plot events.
In the bank robbery scenario, mutual exclusions include pulling a gun
versus handing a note, using a bag versus handling money directly, or
escaping versus being caught.



--
Kurt Luther
www.kurtluther.com

[Kathleen Tuite]

Oh I know (of) that Mark Riedl guy... taking a closer look!

[Robin]

"Narrative intelligence" is his field. Pretty interesting. 

p.s. I'm pretty late but this was an artifact from a cool poster I saw at CHI - http://turnthisinto.com/. In a few of the examples, people are narrating images. 

p.p.s. I start my internship on Monday so will most likely miss the meeting. Will contribute towards notes, though.

Thanks,
Robin
ROBIN N. BREWER
Human-Centered Computing

University of Maryland, Baltimore County

[Yotam Gingold]

In advance of the meeting today, my summary of the "Story Generation with Crowdsourced Plot Graphs" paper by Boyang Li, Stephen Lee-Urban, George Johnston, and Mark O. Riedl.

>>>

They crowdsource simply told stories about a topic (a bank robbery), and analyze (cluster sentences with manual or preliminary crowd cleanup and analyze events' mutually exclusivity and precedence relationships, aggregate by sampling from inputs to create new, remixed stories with a parameter that chooses the typicality of the story. They measure story quality by showing people the stories and letting them edit them; fewer edits is interpreted to mean better stories. The stories they generate are much longer than the original human stories and get edited more (more things are deleted and moved; 68% of the moves are because of inconsistently ordered events).

<<<

Kurt and I have a bunch of stuff to show, and soon, to evaluate.

Yotam

[Kanya(Pao) Siangliulue]

Hi everyone,

I might have to miss the meeting today. I will catch up with Krzysztof later if that is the case.

Cheers,

Pao

[Krzysztof Gajos]

I really like that they capture a reasonable sample of the design space of possible narratives (or a distribution over possible plots if you are a machine learning person or a quantum physicist).  Clearly useful in the context of a training tool, but also a neat trick for folks like us: their representation allows them to simultaneously remove duplicates and create event sequences that none of the individual contributors reported.  Fun!

K.

[Yotam Gingold]

If Pao is going to miss, who will initiate the Google Hangout!???

[Robin]

I forgot I just started an internship in a different time zone. Are there notes about the call somewhere?

Thanks,
Robin
ROBIN N. BREWER
Human-Centered Computing

University of Maryland, Baltimore County

[Yotam Gingold]

Hi Robin,

We discussed the one-paragraph summary of stories we had so-far gathered, and decided to restart it with different parameters (more money). We also added a fourth variant as a control. That has now completed.

We also discussed the impending CSCW deadline (May 31) and tried to volunteer Kurt to outline the paper for the rest of us to fill in.

Yotam

[Robin]

thanks Yotam,

The CSCW deadline is in 7 days! Are we meeting over the weekend or this upcoming Monday to plan the paper?

And are we each writing sections about the components we created?

And are there papers we need to begin aggregating for some sort of lit review section?

apologies for the bombardment of questions, but this deadline is quickly approaching! Perhaps there's a backup conference to submit to in case this doesn't happen on time?

Thanks,
Robin
ROBIN N. BREWER
Human-Centered Computing

University of Maryland, Baltimore County

[Yotam Gingold]

Right! We are meeting on Tuesday at 3pm. We tried to volunteer Kurt to outline the paper. We will write it using google docs.

---
Typed on a tiny keyboard.