10:00 AM - 10:20 AM
[3H1-GS-3d-04] Towards construction of daily living episode ontology
Approach by reusing “activity and participation” categories in ICF
Keywords:Daily living episode, Ontology
There is research related to commonsense reasoning and robotic agents in a human living environment these days. Handling contexts of human daily living is crucial for the systems e.g. detection of risky situations in a care facility and realizing a long-term natural dialogue system that considers the context of daily living.
This research aims to construct a daily living episode ontology to represent daily living activities by reusing the International Classification of Functioning, Disability, and Health (ICF). This ontology provides the common vocabulary for annotating the context to the sensed data in the daily living environment. We extracted the categories from ICF and constructed the is-a hierarchy.
The constructed ontology contains 284 classes. The evaluation of its coverage is performed by using the extracted samples from two resources of commonsense reasoning, i.e. ATlas Of MachIne Commonsense (ATOMIC) and STAIR Actions captions. We tried to describe the context of each sample retrieved from the resources by using the constructed ontology as an evaluation. The results are 0.63 and 0.84 coverage respectively. A detailed discussion will be written in the paper.
This research aims to construct a daily living episode ontology to represent daily living activities by reusing the International Classification of Functioning, Disability, and Health (ICF). This ontology provides the common vocabulary for annotating the context to the sensed data in the daily living environment. We extracted the categories from ICF and constructed the is-a hierarchy.
The constructed ontology contains 284 classes. The evaluation of its coverage is performed by using the extracted samples from two resources of commonsense reasoning, i.e. ATlas Of MachIne Commonsense (ATOMIC) and STAIR Actions captions. We tried to describe the context of each sample retrieved from the resources by using the constructed ontology as an evaluation. The results are 0.63 and 0.84 coverage respectively. A detailed discussion will be written in the paper.
Authentication for paper PDF access
A password is required to view paper PDFs. If you are a registered participant, please log on the site from Participant Log In.
You could view the PDF with entering the PDF viewing password bellow.