Automatically classifying user engagement for dynamic multi-party human–robot interaction
- Submitting institution
-
University of Glasgow
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 11-04509
- Type
- D - Journal article
- DOI
-
10.1007/s12369-017-0414-y
- Title of journal
- International Journal of Social Robotics
- Article number
- -
- First page
- 659
- Volume
- 9
- Issue
- 5
- ISSN
- 1875-4791
- Open access status
- Compliant
- Month of publication
- November
- Year of publication
- 2017
- URL
-
http://eprints.gla.ac.uk/145125/
- Supplementary information
-
-
- Request cross-referral to
- -
- Output has been delayed by COVID-19
- No
- COVID-19 affected output statement
- -
- Forensic science
- No
- Criminology
- No
- Interdisciplinary
- No
- Number of additional authors
-
2
- Research group(s)
-
-
- Citation count
- 16
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- ORIGINALITY: Novel results show a simple hand-written rule, based on sociolinguistic observations of human behaviour, outperformed a set of state-of-the-art machine learning classifiers on engagement detection for a social robot, a crucial task for multi-path embodied interaction, substantially reducing computational complexity and cost. SIGNIFICANCE: IJSR is a premier social robotics journal. Paper combines and extends work from conference papers which have a total of 108 citations. RIGOUR: Five detailed experiments conducted with robot and human participants in real scenarios comparing ML and rule-based classifier. Results showed that online, run-time evaluation is crucial for building successful classifiers.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -