Artificial intelligence and knowledge engineering

Marie-Hélène Abel is a full pro­fes­sor at UTC and, since 2020, is direc­tor of the Depart­ment of com­put­er engi­neer­ing (UTC-GI). Spe­cial­ized in the field of arti­fi­cial intel­li­gence (AI), and more pre­cise­ly, in knowl­edge engi­neer­ing, she gave an inau­gur­al lec­ture on 1st Sep­tem­ber 2020.

After suc­cess­ful­ly defend­ing her doc­tor­al the­sis in AI, in 1994 at UTC, more pre­cise­ly, on “explana­to­ry expert sys­tems”, she was recruit­ed by the Uni­ver­si­ty of Picardie Jules- Verne (UPJV) as a lec­tur­er, then, in 2000, moved to UTC. 

What exact­ly is “an “explana­to­ry expert sys­tem”? It is a sys­tem to which one pro­vides a cer­tain amount of infor­ma­tion, linked to a giv­en prob­lem, from which it will imple­ment spe­cif­ic rea­son­ing and come up a set of var­ied solu­tions. It will also be able to explain its approach to solv­ing the prob­lem so that the user can accept or reject the pro­posed solu­tion,» she explains. 

Her areas of research? «I am inter­est­ed in AI and, more specif­i­cal­ly, in knowl­edge engi­neer­ing, i.e., the work of tool-mak­ing of knowl­edge. Since knowl­edge is spe­cif­ic to each per­son, it may seem sur­pris­ing to work on this theme via tools and machines. How­ev­er, in the field of AI, the aim is to ensure that humans and machines can com­mu­ni­cate and that humans enjoy a sense of mutu­al under­stand­ing. The machine gives an answer that is con­sid­ered rel­e­vant by the human who made the request. In oth­er words, the idea is to use a lan­guage that it is able to read and inter­pret in the same way that a human can,» she says. 

These are areas that require increas­ing­ly sophis­ti­cat­ed algo­rithms. «An algo­rithm is a sequence of instruc­tions that must be read­able and inter­pretable by the machine so that it knows what action to take. Take the High­way Code, for exam­ple. It is a sum of infor­ma­tion from which we all derive the same knowl­edge. A Stop sign is thus inter­pret­ed by all as demand­ing a vehi­cle stop. In the case of algo­rithms, we can speak of sym­bol­ic manip­u­la­tions car­ried out by the machine. Manip­u­la­tions that make sense to the user,» she adds. 

Your mes­sage to stu­dents? «That they should not be afraid of AI because we are often afraid of what we do not know. You have to seize the tools of AI, take advan­tage of all its poten­tial while admit­ting its inher­ent lim­its and risks. We must there­fore be vig­i­lant and impose an eth­i­cal frame­work on our­selves,» con­cludes Marie-Hélène Abel

Le magazine

Novembre 2023 - N°61

Activité physique, nutrition & santé

linkedin facebook pinterest youtube rss twitter instagram facebook-blank rss-blank linkedin-blank pinterest youtube twitter instagram