Séminaire de Dr. LONG Tran-Thanh, University of Southampton, U.K. - Date : mardi 24 mai 2016, 10h00 - Lieu : salle "seminar", Institut MICA, Hanoi University of Science and Technology


Intervenant :
Dr LONG Tran-Thanh, Lecturer in Computer Science, University of Southampton, UK


Date : mardi 24 mai 2016, 10h00
Lieu : "seminar room", 9ème étage, bâtiment B1, Institut MICA, Hanoi University of Science and Technology
Langue: le séminaire sera présenté en anglais


With the recent fantastic breakthroughs in Artificial Intelligence (AI), such as the latest success of AlphaGo or the advancements in robotics, comes along an increasing number of concerns about the dangers and threats these Ai technologies may bring to our society. These concerns may become so serious in the future that it would cause serious harms to further advancements of AI research. In fact, the root of these concerns lies within the fear of creating a superhuman Artificial General Intelligence (AGI) that one day may decide to destroy the humankind. To overcome these concerns, there have been many attempts to position AI as a set of more human-friendly and less threatening technologies.
A very promising direction of these attempts is the concept of collaborative AI. This concept significantly differs from the AGI approach, as instead of focusing on creating superhuman competitors, it still keeps the human factor at the centre of its objectives. In particular, collaborative AI provides technologies that aim to ease our everyday life in a supportive and ubiquitous way. As ubiquitous systems, such as Internet of Things, and their applications (e.g., smart homes, smart cars, or smart cities) are becoming more and more successful, I argue that collaborative AI will also become a dominant concept in the (very) near future.
However, state-of-the-art collaborative AI is still in its infant stage, and it will have to overcome a number of obstacles in order to achieve maturity. As such, in this talk, I will first describe in detail a number of major obstacles of the concept, namely: (i) human participation motivation; (ii) flexible autonomy; (iii) agile teaming; (iv) user privacy; and (v) cyber security. In the second part of the talk, I will discuss the state-of-the-art research solutions within each abovementioned topic. In particular, I will mainly focus on the problem of having the human-factor in optimisation problems, a research area I have been working on with my collaborators from Southampton, Oxford, and University of Southern California.