Computing With Perceptions | Mohamed Quafafou15h30 | Salle du conseil de Polytech Angers

Le 26 novembre 2021

Computing With Perceptions

Mohamed Quafafou,
LIS CNRS, Aix-Marseille University, France.
mohamed.quafafou@univ-amu.fr

Abstract:
Today, more than ever, machines are going beyond their usual role, performing many numerical calculus operations. Machines play a human-like role in everyday life, just think of Chatbots that conduct an online textual conversation with a human, humanoid robots that accompany elderly persons, intelligent avatars used in e-commerce, etc. But can machines have the abilities of humans to live in the world? Humans achieve goals during their daily life using, among others, their ability to think. Hence, can machines think? Alain Turing considered this question in his paper entitled “Computing Machinery and Intelligence” underlying that the words “machines” and “think” are ambiguous. Thus, he replaces the question with the “imitation game” problem based on a question/answer method. This method seems to be suitable as it reflects humans’ attempts to achieve a goal. On its part, cognitive science, which is the interdisciplinary study of mind and intelligence, goes beyond computation by assuming that thinking is representational and computational. Hence, cognitive science considers mental procedures, which operate on structures or mental representations like logical propositions, rules, concepts, images, and analogies. Also, cognitive anthropology examines human thinking in the light of the cultural dimension. However, cognitive science turns a blind eye to the role of the physical world in human thinking while human perceptions affect reasoning and decision-making. As pointed by John McCarthy, today’s machines or robots are limited because their relation to the world is almost non-existent: “What the robot believes about the world, in general, doesn’t arise for the limited robots of today, because the languages they are programmed to use can’t express assertions about the world in general”. Humans manage the aspects of everyday life by allowing or tolerating contradictions and ambiguities, whereas machines perform computation under consistency (no contradiction) and completeness (no ambiguity) constraints. Moreover, the perception is influenced by several other factors like education, culture, and past experiences.

In this talk, we give an overview of AI-Human collaboration. This collaboration is crucial as machines are playing human-like roles and will work with humans in hybrid teams. Moreover, the nature of humans and machines are very different: humans perceive the world through their five senses and act accordingly, whereas machines perform digital computations by running programs. One of the aims of computing with perception is to enhance the ability of AI to collaborate with humans in order to solve problems in which decisions are based on both measurements and perceptions.

Mohamed Quafafou  is a professor of computer science at Aix-Marseille University. His main research interests are in Data Mining theory and applications. Prior to joining Aix-Marseille University in 2005, he served as professor at Avignon University from 2002 to 2004 and as Assistant Professor at University of Nantes from 1995 to 2001. Mohamed Quafafou received his Habilitation Ph.D. in 2000 on Rough Set Theory and Data Mining, and his Ph.D. in Computer Science from INSA Lyon, in 1992. He received an M.Sc. in computer science from Paul Sabatier University (Toulouse). In partnership with France Telecom, he developed the first French web mining system discovering emergent web communities. He is the head of the research group on “data mining at scale” at LIS-CNRS and computer science department at Polytech Marseille.