SCHOENAUER Marc
Faculty habilitation
Group : Learning and Optimization

SCHOENAUER.01-01-1997

Starts on
Advisor :

Funding :
Affiliation : Université Paris-Saclay
Laboratory :

Defended on 01/06/1997, committee :

Research activities :
   - Artificial Intelligence
   - Evolutionary computation
   - Stochastic optimization

Abstract :
The work described in this document is centered on Evolutionary Computation and some of its applications in different fields of numerical optimization.

I graduated in Numerical Analysis, and my trajectory toward Computer Science and Evolutionary Computation is a little sneaky: After a smooth standard start with my PhD thesis in June 1980, in the domain of Numerical Analysis, I began to be interested in the potential applications of AI to that field. Starting from Expert Systems, and due to the frequent absence of experts in such technical domains, we soon turned to Machine Learning All my work in the Machine Learning area is joint work with M. Sebag, LMS (Solid Mechanics Laboratory, Ecole Polytechnique) and LRI (Computer Science Dpt, University Paris XI) with substancial help from Prof. J. Zarka, LMS.

But rapidly, after getting answers to the direct question: ``If feature A is a and feature B is b, then what happens?'' (Answer from the rule base: ``Catastrophe C is likely to happen'' ), the experts asked the inverse question : ``How can I tune features A and B to avoid catastrophe C?''

Such inverse problems are amenable to the field of Optimization, but standard tools are of poor use in semi-discrete spaces, with very rough objective functions: At that time, I was ready to meet Evolutionary Computation: the encounter with Genetic Algorithms happened in the late 80's.

The main trends of my work now are Function Identification and Structure Optimization on the application side, and hybridization of evolutionary techniques, numerical analysis methods and inductive learning algorithms on the more fundamental side.

I am mostly interested in Evolutionary Algorithms as Function Optimizers, and I feel Evolutionary Computation should be part of the standard optimization tool boxes of Numerical Analysis. On the other hand, my background in Numerical Analysis certainly is an asset for somebody working in the field of Optimization. Nevertheless, I am really convinced both fields can benefit from their civilized meeting.

The present document summarizes past, present and future work of EEAAX, the Artificial Evolution and Machine Learning Team I am leading at Ecole Polytechnique (Equipe Evolution Artificielle et Apprentissage de l'X), as this is the part of my scientific work for which I ask to be habilite a diriger des recherches.

November 1995.

CAUSAL LEARNING FOR DIAGNOSTIC SUPPORT


CAUSAL UNCERTAINTY QUANTIFICATION UNDER PARTIAL KNOWLEDGE AND LOW DATA REGIMES


MICRO VISUALIZATIONS: DESIGN AND ANALYSIS OF VISUALIZATIONS FOR SMALL DISPLAY SPACES
The topic of this habilitation is the study of very small data visualizations, micro visualizations, in display contexts that can only dedicate minimal rendering space for data representations. For several years, together with my collaborators, I have been studying human perception, interaction, and analysis with micro visualizations in multiple contexts. In this document I bring together three of my research streams related to micro visualizations: data glyphs, where my joint research focused on studying the perception of small-multiple micro visualizations, word-scale visualizations, where my joint research focused on small visualizations embedded in text-documents, and small mobile data visualizations for smartwatches or fitness trackers. I consider these types of small visualizations together under the umbrella term ``micro visualizations.'' Micro visualizations are useful in multiple visualization contexts and I have been working towards a better understanding of the complexities involved in designing and using micro visualizations. Here, I define the term micro visualization, summarize my own and other past research and design guidelines and outline several design spaces for different types of micro visualizations based on some of the work I was involved in since my PhD.