Project

Back to overview

CrowdDB: Integrating Human Input into Databases

English title CrowdDB: Integrating Human Input into Databases
Applicant Kossmann Donald
Number 137613
Funding scheme Project funding
Research institution Institut für Informationssysteme ETH Zürich
Institution of higher education ETH Zurich - ETHZ
Main discipline Information Technology
Start/End 01.12.2011 - 30.06.2015
Approved amount 305'964.00
Show all

Keywords (6)

database systems; query processing; crowd-sourcing; human computation; automatic user interface generation; open world data models

Lay Summary (English)

Lead
Lay summary
Database technology has matured over the last twenty to thirty years.  As a result, database systems have achieved widespread adoption in many application domains such as the management of business data, personal data, or scientific data.  Despite of this success, there are still a number of hard data management problems that current database systems do not handle well. Examples include scenarios in which data is missing, data is redundant (e.g., the same entity is represented in different ways multiple times), or subjective comparisons need to be carried out (e.g., identifying objects on images).  While current database systems are not well suited to address these challenges, these challenges can be easily addressed by humans.  The goal of this project is to develop new techniques to extend modern database systems such that they can integrate human input.  This way, complex tasks and data management scenarios can be addressed by letting machines do what machines are best for (e.g., number crunching, scanning through large volumes of data) and by letting humans do what humans are best for (e.g., detecting and resolving inconsistencies).
Direct link to Lay Summary Last update: 21.02.2013

Responsible applicant and co-applicants

Employees

Publications

Publication
Cost-Efficient Querying Strategies for the Crowd
Grünheid Anja, Nushi Besmira, Kossmann Donald (2014), Cost-Efficient Querying Strategies for the Crowd, in Proceedings of the Big Uncertain Data Workshop, SIGMOD 2014 .
Incremental Record Linkage
Gruenheid A., Dong X.L., Srivastava D. (2014), Incremental Record Linkage, in Proceedings of the VLDB Endowment, Vol. 7 (9), 2014.
Quality Assurance and Crowd Access Optimization: Why does diversity matter?
Nushi Besmira, Singla Adish, Gruenheid Anja, Krause Andreas, Kossmann Donald (2014), Quality Assurance and Crowd Access Optimization: Why does diversity matter?, in Proceedings of the ICML Workshop on Crowdsourcing and Human Computing 2014 .
Cost and Quality Trade-Offs in Crowdsourcing
Grünheid Anja, Kossmann Donald (2013), Cost and Quality Trade-Offs in Crowdsourcing, in VLDB Workshop on Databases and Crowdsourcing, Trento, Italy.
When is A=B?
Grünheid Anja, Kossmann Donald, Nushi Besmira (2013), When is A=B?, in EATCS Bulletin, (111), 88-97.
CrowdDB: answering queries with crowdsourcing
Franklin Michael J., Kossmann Donald, Kraska Tim, Ramesh Sukriti, Xin Reynold (2011), CrowdDB: answering queries with crowdsourcing, in In Proceedings of the 2011 ACM SIGMOD International Conference on Management of data, pp. 61-72. ACM.
Crowddb: Query processing with the vldb crowd
Feng Amber, Franklin Michael, Kossmann Donald, Kraska Tim, Madden Samuel, Ramesh Sukriti, Wang Andrew, Xin Reynold (2011), Crowddb: Query processing with the vldb crowd, in Proceedings of the VLDB Endowment 4, no. 12 (2011). Demo.
Crowd Access Path Optimization: Diversity Matters
Nushi B., Singla A., Gruenheid A., Zamanian E., Krause A., Kossmann D., Crowd Access Path Optimization: Diversity Matters, in Proceedings of the Conference on Human Computation & Crowdsourcing (HCOMP) 2015 .
CrowdSTAR: A Social Task Routing Framework for Online Communities
Nushi Besmira, Alonso Omar, Hentschel Martin, Kandylas Vasileios, CrowdSTAR: A Social Task Routing Framework for Online Communities, in Proceedings of the ICWE 2015., Rotterdam, NL.

Collaboration

Group / person Country
Types of collaboration
University of California, Berkeley United States of America (North America)
- in-depth/constructive exchanges on approaches, methods or results
- Publication
- Research Infrastructure
Brown University United States of America (North America)
- in-depth/constructive exchanges on approaches, methods or results
- Publication
Massachusetts Institute of Technology United States of America (North America)
- in-depth/constructive exchanges on approaches, methods or results
Carnegie Mellon University United States of America (North America)
- in-depth/constructive exchanges on approaches, methods or results

Awards

Title Year
Fritz -Kutter Award for Sukriti Ramesh for her master thesis: CrowdDB – Answering Queries with Crowdsourcing http://www.kutter-fonds.ethz.ch/PreviousPrizewinners.html 2011

Abstract

Database technology hasmaturedover the last twenty tothrirtyyears.Despitethese success stories, thereare still some notoriously hardresearch problems. More specifically, thereare still some queries that state-of-the-art database and information systems such as Oracleor Google cannot answer. Processing these queries involves intelligence in order to deal with inconsistencies or missing data. This research project will explore how to integrate human input effectively in order to process such queries.Specifically, the plan is to develop a new kindof database management system that decomposes queries such that parts of the query are executed by machines and other parts are executed by humans. This way, machines carry out tasks that machines can do well (e.g., numbercrunching, processing large volumes of data) and humansare involvedonly intasks that machines cannot handle well (e.g., resolve inconsistencies and enter missing information). Building such a systemraises many technical research challenges in terms of query language and semantics, optimization, usability, and privacy. We planto address these technical challenges in this project. Buildingsuch a system also raises a number of other challenges; e.g., legal,social,and business. Studying these other challenges is beyond the scope ofthis project.
-