QALD-3 » Program
September 2013 ⋅ Co-located with: CLEF 2013  

Workshop Program

Day: September 25 (Wednesday), 2013

 
14:30 Welcome and Overview of QALD-3
 
15:00 Invited Talk: Anthony Fader (University of Washington)
Open-Domain Question Answering
 
16:00 Cristina Giannone (University of Rome Tor Vergata)
A HMM-based Approach to Question Answering against Linked Data
 
16:30 Coffee break
 
17:00 Camille Pradel (IRIT, Université Paul Sabatier, Toulouse)
SWIP at QALD-3: results, criticisms and lesson learned
 
17:30 Corina Dima (Universität Tübingen)
Intui2: A Prototype System for Question Answering over Linked Data
 
18:00 Open Discussion on the Future of QALD

 

Invited Talk: Anthony Fader (University of Washington, homepage)
Open-Domain Question Answering

Creating general question-answering (QA) systems is a long-standing problem in artificial intelligence. QA systems that interpret questions, reason over evidence, and return useful answers would be a powerful tool for users, potentially replacing search engines as the de facto standard for information access on the web.

In this talk, I will present some recent work towards creating a large-scale, open-domain QA system that (1) automatically constructs a knowledge base from text, and (2) learns to answer questions over the knowledge base using social QA sites (e.g. WikiAnswers) as a source of supervision. The first part of my talk will focus on a novel algorithm that learns from naturally occurring question-paraphrases, without requiring any manually annotated questions. The second part of my talk will focus on scaling the system to work with larger knowledge bases, handle more complex questions, and rank answers for higher precision.