AI Strategy and Concepts Bibliography

WIKINDX Resources

Amir, O. (2017). Intelligent information sharing to support loosely-coupled teamwork. Unpublished PhD thesis. 
Added by: SijanLibrarian (2021-09-14 16:09:47)   Last edited by: SijanLibrarian (2021-09-14 16:12:22)
Resource type: Thesis/Dissertation
BibTeX citation key: Amir2017
Email resource to friend
View all bibliographic details
Categories: Artificial Intelligence, Cognitive Science, Complexity Science, Computer Science, Data Sciences, Decision Theory, General
Subcategories: Behavioral analytics, Big data, Deep learning, Informatics, Machine learning, Psychology of human-AI interaction
Creators: Amir
Views: 32/40
Views index: 18%
Popularity index: 4.5%

Complex tasks such as treating patients with complex medical conditions, conducting research, co-authoring documents and developing software products are typically accomplished by teams. Teamwork in such settings is often loosely-coupled as team members assume different responsibilities that match their individual expertise. This decomposition of activities enables team members to function autonomously and requires team members to be aware of others’ actions only if these actions interact with their own activities. However, identifying interactions between collaborators’ activities can be challenging. As a result, team members are often overwhelmed by too much irrelevant information about others’ activities, or lack important relevant information, both of which can lead to coordination failures. This thesis argues that intelligent information sharing methods that identify the information that is most relevant to each team member can reduce coordination overhead and improve team performance.

Through a study of teams caring for children with complex medical conditions, this thesis characterizes the coordination challenges of loosely-coupled teams and formalizes the problem of information sharing in such teams. The thesis introduces Mutual Influence Potential Networks, a new representation for modeling collaborative activities. It further defines MIP-DOI, an algorithm which uses the Mutual Influence Potential Network representation to identify the most relevant information for each team member.

The thesis also presents the design, implementation and evaluation of a personalized change awareness mechanism, which uses MIP-DOI to reduce the amount of shared change information in the context of collaborative writing. The results of an experiment evaluating this mechanism show that compared to the currently most prevalent approach of presenting users with all changes made by their collaborators, the personalized change awareness mechanism resulted in significantly reduced perceived workload and significantly increased productivity of team members. Importantly, the personalized change awareness mechanism did not have any detrimental effect on the quality of the work.

wikindx 6.2.2 ©2003-2020 | Total resources: 1447 | Username: -- | Bibliography: WIKINDX Master Bibliography | Style: American Psychological Association (APA) | Database queries: 56 | DB execution: 0.13989 secs | Script execution: 0.15264 secs