TaskGenie: Crowd-Powered Task Generation for Struggling Search

Luyan Xu, Xuan Zhou, Ujwal Gadiraju

Research output: Chapter in Book/Conference proceedings/Edited volumeConference contributionScientificpeer-review

1 Citation (Scopus)
25 Downloads (Pure)


Search tasks provide a medium for the evaluation of system performance and the underlying analytical aspects of IR systems. Researchers have recently developed new interfaces or mechanisms to support vague information needs and struggling search. However, little attention has been paid to the generation of a unified task set for evaluation and comparison of search engine improvements for struggling search. Generation of such tasks is inherently difficult, as each task is supposed to trigger struggling and exploring user behavior rather than simple search behavior. Moreover, the everchanging landscape of information needs would render old task sets less ideal if not unusable for system evaluation. In this paper, we propose a task generation method and develop a crowd-powered platform called TaskGenie to generate struggling search tasks online. Our experiments and analysis show that the generated tasks are qualified to emulate struggling search behaviors consisting of ‘repeated similar queries’ and ‘quick-back clicks’, etc. – tasks of diverse topics, high quality and difficulty can be created using this framework. For the benefit of the community, we publicly released the platform, a task set containing 80 topically diverse struggling search tasks generated and examined in this work, and the corresponding anonymized user behavior logs.

Original languageEnglish
Title of host publicationWeb Information Systems Engineering
Subtitle of host publicationWISE 2020 - Proceedings
EditorsZhisheng Huang, Wouter Beek, Hua Wang, Yanchun Zhang, Rui Zhou
Place of PublicationCham
Number of pages18
ISBN (Electronic)978-3-030-62008-0
ISBN (Print)978-3-030-62007-3
Publication statusPublished - 2020
Event21st International Conference on Web Information Systems Engineering, WISE 2020 - Amsterdam, Netherlands
Duration: 20 Oct 202024 Oct 2020

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


Conference21st International Conference on Web Information Systems Engineering, WISE 2020

Bibliographical note

Green Open Access added to TU Delft Institutional Repository ‘You share, we take care!’ – Taverne project https://www.openaccess.nl/en/you-share-we-take-care
Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.


  • SERPs
  • User behavior
  • User interaction
  • Web search


Dive into the research topics of 'TaskGenie: Crowd-Powered Task Generation for Struggling Search'. Together they form a unique fingerprint.

Cite this