Intelligent Personal Assistants (IPAs) such as Apple Siri, Google Home, Amazon Alexa etc play
a significant role in our everyday lives by assisting us in the information search and other day to
day tasks. These systems, however, possess the reactive behaviour which is achieved through a
request-response approach. While this behaviour is still helpful for the users to some extent, it
does not reduce the actions necessary to perform a particular task and the cognitive load in
managing them, which eventually leads to the users’ exerting a great amount of effort to achieve
those tasks [1]. Furthermore, the users’ goals are not sufficiently modeled in these systems and
as a result, a reference model of IPAs proactively acting on users’ behalf can be found missing.
In this project, we want to take the IPAs research to the next level by integrating the users’ goals
in addition to the proactive behaviour. Detecting and considering the users’ goals or intentions
will help these systems in goal-based decision making and delivering intervention or providing the
support proactively [2]. The rapid development of artificial intelligence (AI) technology has enabled the multi-agent modeling and simulation technology of complex systems to advance rapidly. These AI technologies for example the multi-agent reinforcement learning modeling and
simulation [3] would be leveraged to study the behaviour of the proposed proactively personalised
IPAs.
This research project will examine a simulation framework for Proactive IPAs and will be aligned with research from a team in the ADAPT Centre that is examining several challenges related to Proactive Intelligent Personal Assistants.
References
[1] Myers, Karen & Yorke-Smith, Neil. (2008). Proactive behavior of a personal assistive agent.
[2] Meurisch, Christian & Ionescu, Maria-Dorina & Schmidt, Benedikt & Mühlhäuser, Max.
(2017). Reference model of next-generation digital personal assistant: integrating proactive
behavior. 149-152. 10.1145/3123024.3123145.
[3] W. Fan, P. Chen, D. Shi, X. Guo and L. Kou, “Multi-agent modeling and simulation in the AI
age,” in Tsinghua Science and Technology, vol. 26, no. 5, pp. 608-624, Oct. 2021, doi:
10.26599/TST.2021.9010005.