Human computation

Human Computation or human-based computation (roughly translated: human-based information processing ) describes a paradigm in computer science in which a large number of human users solve tasks that are difficult or impossible to solve by machine. These tasks include many language and image processing problems in which Artificial Intelligence does not reach the level of performance of humans.

The philosopher Rainer Mühlhoff coined the term "human-assisted artificial intelligence" to point out that many deep learning applications in artificial intelligence are based on human cooperation, as they continuously rely on training and verification data. For example, through click work , gamification or CAPTCHAs , people are integrated into hybrid computer networks, the calculations of which are then partially carried out in human brains and on silicon-based processors .

See also

• Data science
• Crowdsourcing
• Language processing
• Image processing
• Citizen Science

swell

1. ↑ http://dm.tzi.de/human-computation/
2. ↑ Rainer Mühlhoff: Human- assisted artificial intelligence: About the socio-technical requirements of "deep learning" . In: ZfM - Journal for Media Studies . tape 21 , no. 2 . transcript, 2019, ISSN  1869-1722 , p. 56-64 ( philpapers.org ).
3. ↑ Rainer Mühlhoff: Human-aided artificial intelligence: Or, how to run large computations in human brains? Toward a media sociology of machine learning . In: New Media & Society . November 6, 2019, ISSN  1461-4448 , p. 1–17 , doi : 10.1177 / 1461444819885334 ( sagepub.com [accessed November 22, 2019]).

literature

• Michelucci, Pietro (Ed.): Handbook of Human Computation . Springer, 2013, ISBN 978-1-4614-8805-7 .