Help us map answer lines to Wikipedia pages!

Packet databases and other quizbowl sites, apps, or software should be discussed here.
Post Reply
User avatar
ezubaric
Rikku
Posts: 369
Joined: Mon Feb 09, 2004 8:02 pm
Location: College Park, MD
Contact:

Help us map answer lines to Wikipedia pages!

Post by ezubaric »

As has become tradition, we’re doing another exposition match of our quiz bowl system QANTA against quiz bowlers at HSNCT in Atlanta.

A big part of our system is having good data. In the past, we used automatic method to map answer lines to concepts in our knowledge base. However, we’re realizing that this is incomplete and introduced some errors. We’d like the community’s help in cleaning up some of our data.

We’re going to be working on it ourselves, but we’d appreciate help from anyone who has a few moments to help out. It’s kinda addictive and helps expose you to answer lines that you might otherwise not know about. (In other words, you can use it as a study mechanism.)

The basic idea is that we want to map answer lines to Wikipedia pages. Below is a link to a Google Spreadsheet where we’re collecting mappings from answer lines to Wikipedia pages. Our primary goal is to move as many lines from _Automatic Guesses_ sheet to the _Unambiguous_ sheet. We also have sheets for ambiguous assignments and for impossible assignments, but those are less important at the moment.

Many of the automatic guesses will be correct, but we’d like human eyes to verify all of the assignments. I’d suggest a strategy of taking a chunk of answers from the automatic assignment, copy them over to ambiguous and then deleting / correcting the wrong entries (if you move things to the ambiguous sheet, that’s great too, but unambiguous is the higher priority right now).

Don’t worry about blank lines in the sheet. We’ll periodically do a cleanup to make sure that it’s not too messy.

If you have questions about specific entries, please use the built-in commenting feature, as that will bring us directly to the row you have questions about.

Google Spreadsheet:
https://docs.google.com/spreadsheets/d/ ... sp=sharing

Document Describing our System (go to the end for page assignment information):
https://github.com/Pinafore/qb/blob/ingestion/README.md

If you have questions you don't want to ask here, you can e-mail me at [email protected]

If you're really into this (it can get addictive), we can add you to the slack channel. If you're willing to commit to doing a lot of these, we can probably get an Amazon gift card for you.
Last edited by ezubaric on Tue May 09, 2017 5:38 pm, edited 2 times in total.
Jordan Boyd-Graber
UMD (College Park, MD), Faculty Advisor 2018-present
UC Boulder, Founder / Faculty Advisor 2014-2017
UMD (College Park, MD), Faculty Advisor 2010-2014
Princeton, Player 2004-2009
Caltech (Pasadena, CA), Player / President 2000-2004
Ark Math & Science (Hot Springs, AR), Player 1998-2000
Monticello High School, Player 1997-1998

Human-Computer Question Answering:
http://qanta.org/
User avatar
Skepticism and Animal Feed
Auron
Posts: 3238
Joined: Sat Oct 30, 2004 11:47 pm
Location: Arlington, VA

Re: Help us map answer lines to Wikipedia pages!

Post by Skepticism and Animal Feed »

Wait I don't get it. So right now for example "Caesar Rodney" is listed as an answerline, and "Julius_Caesar" is listed as the wikipedia article. Should I be deleting the entire row, or should I be changing the Wikipedia article to Caesar_Rodney?
Bruce
Harvard '10 / UChicago '07 / Roycemore School '04
ACF Member emeritus
My guide to using Wikipedia as a question source
User avatar
ezubaric
Rikku
Posts: 369
Joined: Mon Feb 09, 2004 8:02 pm
Location: College Park, MD
Contact:

Re: Help us map answer lines to Wikipedia pages!

Post by ezubaric »

At a minimum, please change the Wikipedia page to the correct one. If you also want to move that over to the "unambiguous" tab, that would be great.

Thanks!
Jordan Boyd-Graber
UMD (College Park, MD), Faculty Advisor 2018-present
UC Boulder, Founder / Faculty Advisor 2014-2017
UMD (College Park, MD), Faculty Advisor 2010-2014
Princeton, Player 2004-2009
Caltech (Pasadena, CA), Player / President 2000-2004
Ark Math & Science (Hot Springs, AR), Player 1998-2000
Monticello High School, Player 1997-1998

Human-Computer Question Answering:
http://qanta.org/
User avatar
ezubaric
Rikku
Posts: 369
Joined: Mon Feb 09, 2004 8:02 pm
Location: College Park, MD
Contact:

Re: Help us map answer lines to Wikipedia pages!

Post by ezubaric »

We don't want to delete rows without a correct Wikipedia assignment, as those answer lines still need to get taken care of at some point.
Jordan Boyd-Graber
UMD (College Park, MD), Faculty Advisor 2018-present
UC Boulder, Founder / Faculty Advisor 2014-2017
UMD (College Park, MD), Faculty Advisor 2010-2014
Princeton, Player 2004-2009
Caltech (Pasadena, CA), Player / President 2000-2004
Ark Math & Science (Hot Springs, AR), Player 1998-2000
Monticello High School, Player 1997-1998

Human-Computer Question Answering:
http://qanta.org/
A Dim-Witted Saboteur
Yuna
Posts: 973
Joined: Tue Aug 02, 2016 12:31 pm
Location: Indiana

Re: Help us map answer lines to Wikipedia pages!

Post by A Dim-Witted Saboteur »

Sorry if this is posted elsewhere, but how/where do I sign up to play this robot?
Jakob M. (they/them)
Michigan State '21, Indiana '2?
"No one has ever organized a greater effort to get people interested in pretending to play quiz bowl"
-Ankit Aggarwal
User avatar
ezubaric
Rikku
Posts: 369
Joined: Mon Feb 09, 2004 8:02 pm
Location: College Park, MD
Contact:

Re: Help us map answer lines to Wikipedia pages!

Post by ezubaric »

We don't have a way to play it online yet; we've just focused on in-person exhibition matches. We're hoping that we can have an online system that people can play this summer.
Jordan Boyd-Graber
UMD (College Park, MD), Faculty Advisor 2018-present
UC Boulder, Founder / Faculty Advisor 2014-2017
UMD (College Park, MD), Faculty Advisor 2010-2014
Princeton, Player 2004-2009
Caltech (Pasadena, CA), Player / President 2000-2004
Ark Math & Science (Hot Springs, AR), Player 1998-2000
Monticello High School, Player 1997-1998

Human-Computer Question Answering:
http://qanta.org/
User avatar
ezubaric
Rikku
Posts: 369
Joined: Mon Feb 09, 2004 8:02 pm
Location: College Park, MD
Contact:

Re: Help us map answer lines to Wikipedia pages!

Post by ezubaric »

Thanks for all your help! (Including some answer lines that made it into this match!)

Video: https://www.youtube.com/watch?v=bYFqMIN ... e=youtu.be

We're going to focus on more researchy things for a bit, so we're shutting down the answer mapping Google Doc. If somebody out there really has a hankering to do it, please contact me.

(Or if you're a programmer who'd like to take over answer mapping and data ingestion for us, that would be a huge burden lifted off our shoulders.)
Jordan Boyd-Graber
UMD (College Park, MD), Faculty Advisor 2018-present
UC Boulder, Founder / Faculty Advisor 2014-2017
UMD (College Park, MD), Faculty Advisor 2010-2014
Princeton, Player 2004-2009
Caltech (Pasadena, CA), Player / President 2000-2004
Ark Math & Science (Hot Springs, AR), Player 1998-2000
Monticello High School, Player 1997-1998

Human-Computer Question Answering:
http://qanta.org/
Post Reply