Titulo Estágio
Empowering IoT with Intelligent Personal Assistant Agents
Áreas de especialidade
Sistemas Inteligentes
Local do Estágio
DEI-FCTUC
Enquadramento
The increasing deployment of Artificial Intelligence and Machine Learning coupled with the rapid penetration of mobile devices boosted the development of systems towards what has been termed Intelligent Personal Assistant Agents (IPAAs), in which Artificial Intelligence techniques are drawn upon to create systems capable of providing people with more personalized and adaptive assistance (e.g. Apple Siri). IPAAs can monitor the behavior of the individual and produce models of what the individual knows, how they feel, what are their motives, intentions and desires. This user model, i.e, a representation of information about an individual user, is essential for an adaptive IPAA to provide the adaptation effect, i.e., to behave differently for different users Empowered by these user models, IPAAs can predict for specific contexts these mental states of their users and act appropriately
Wearables offfer unprecedent opportunities for tracking persons’ movements, ambient environment, biosignals and much more, all in real time, to monitor, document and augment their lives. Moreover, coupled with advances in IoT and artificial intelligence/machine learning, wearables hold the promise of achieving a new level of human connectedness.
Although considerable work has been done in the recent past regarding IoT, most technologies and solutions for accessing real-world information are either closed, platform-specific, or application-specific. Existing architectures can be classified as device-centric, technology-centric, service-centric or entity-centric, the latter dealing with entities such as goods, cars or people exactly in the same way. So, on one side, there is the need to define a people-centric IoT architecture that, building on existing architectures, goes beyond devices, technology, services and entities and, on the other side, develop a set of common building blocks, middleware and/or services that can be used to construct people-oriented Cyber-Physical Systems (CPS) applications, in domains including but not restricted to mHealth, social networks enhancement, and Ambient-Assisted Living (AAL), in an open and more effective way.
One of the main objectives of the project SOCIALITE (Social-Oriented Internet of Things Architecture, Solutions and Environment. PTDC/EEI-SCR/2072/2014) is to define a first generic version of an IoT architecture that can be used for building specific solutions for each particular people-centric application domain. A distinguishing aspect concerning this project is the focus on people. The idea is that communication between people to people and people to things will be smarter, more intelligent, transparent and unobtrusive to the user, and adapted to the context of the users.
Objetivo
The goal of this project is the incorporation of existing probabilistic graphic models (bayesian belief nets) or other machine learning techniques (naive bayes, etc) in a IPAA so that it can, autonomously or semi-autonomously and in an unsupervised/supervise/semi-supervised way, learn/infer as much as possible the beliefs of the user, his/her motivational tendencies (goals), his/her affective state, given a context described as a set of attributes. Afterwards, the personal agent should be able to make recommendations accordingly.
For instance, in the e-learning domain, the personal agent may acquire information about a student that is frequenting a course at distance (e.g., what the student already knows, what skills and knowledge he/she already has, what courses he/she got success, the performance in specific themes of the course, what are his/her interests, affetive status, mood, emotions) and based on this information make appropriate recommendations in order to improve student’s performance.
Plano de Trabalhos - Semestre 1
1- State of the Art [Sept – Oct]
2- Analysis and Specification [Nov]
3- Definition of System Requirements [Dec]
4- Prototype Development [Dec – Jan]
5- Thesis Proposal Writing [Dec – Jan]
Plano de Trabalhos - Semestre 2
6- Prototype Improvement [Feb – Apr]
7- Experimental Tests [Apr – May]
8- Paper Writing [May – Jun]
9- Thesis Writing [Jun – Jul]
Condições
O trabalho será desenvolvido num laboratório do CMS, com recurso a meios computacionais adequados. O estágio será parcialmente remunerado se o aluno demonstrar desempenho compatível.
Observações
Requisitos:
Competências de Programação
Disciplina de Inteligência Artificial (MEI)
Motivação para projectos práticos de Ciência Cognitiva e Inteligência Artificial
Orientador
Luis Miguel Machado Lopes Macedo
macedo@dei.uc.pt 📩