Project Goals

The idea of the web is to form an open space for collaboration, where the distinction between author and reader fades away. With the current development of Web 2.0, this is actually happening at a broad scale, e.g.:

  • People write blogs.
  • Wikis are in widespread use.
  • Collaborative tagging systems rival professionally built taxonomies.

The users of Web 2.0 now produce huge amounts of information of diverse quality. This leads to the problem of information overload: how to make the most of this information without overwhelming the users? One key challenge to solve this issue is to assess the quality of the user generated content.

In AQUA, we seek to develop algorithms to assess the quality of content automatically. We focus on two sources for this assessment: (1) user generated content ; (2) feedback by users of the content, i.e. user generated meta-data, e,g, in the form of annotations like ratings. To do so, we investigate techniques from the fields of natural language processing (NLP), information retrieval and machine learning.

So, in a nutshell, AQUA will answer the following questions:

  • What is quality of information? How does it matter in information search?
  • How to model the quality of user generated content?
  • How far can you go with automatic methods in assessing quality?
  • How to give feedback to users regarding quality?

The AQUA project is associated with the project "Mining Lexical-Semantic Knowledge from Dynamic and Linguistic Sources and Integration into Question Answering for Discourse-Based Knowledge Acquisition in eLearning(QA-EL)".

Publications related to AQUA

PDF Abstract BibTeX Entry
A Monolingual Tree-based Translation Model for Sentence Simplification. COLING2010, to appear.
PDF Abstract BibTeX Entry
A Multi-Dimensional Model for Assessing the Quality of Answers in Social Q&A Sites (Extension to ICIQ2009). Technical Report TUD-CS-2009-0158, October 2009.
PDF Abstract BibTeX Entry
A Multi-Dimensional Model for Assessing the Quality of Answers in Social Q&A Sites. In:Proceedings of 14th International Conference on Information Quality, November 2009.
PDF Abstract BibTeX Entry
Kateryna Ignatova, Delphine Bernhard, Iryna Gurevych:
Generating High Quality Questions from Low Quality Questions. In: Workshop on the Question Generation Shared Task and Evaluation Challenge, September 2008.
PDF Abstract BibTeX Entry
Predicting the Perceived Quality of Web Forum Posts. In: Proceedings of the Conference on Recent Advances in Natural Language Processing (RANLP), 2007.
PDF Abstract BibTeX Entry
Automatically Assessing the Post Quality in Online Discussions on Software. In: Companion Volume of the 45rd Annual Meeting of the Association for Computational Linguistics (ACL), 2007.
PDF Abstract BibTeX Entry
Notetaking Behavior in University Courses and its Implications on eLearning Systems. In: Tagungsband der fünften e-Learning Fachtagung Informatik, 2007.

Principal Investigator

Former Research Staff

  • Jürgen Steimle (GK "eLearning")
  • Markus Weimer (GK "eLearning")

Project Publications

Additional Attributes

Type

A A A | Drucken Print | Impressum Impressum | Sitemap Sitemap | Suche Search | Kontakt Contact | Webseitenanalyse: Mehr Informationen
zum Seitenanfangzum Seitenanfang