My research has evolved very considerably, in both its subject matter and methods, since my doctoral thesis (Princeton, USA, 1995). I started my research working on theoretical problems in ``early universe'' cosmology i.e. in the very hot dense phase of the universe postulated by cosmological models. I studied notably, and elaborated possible solutions to, the problem of the origin of the observed excess of particles of matter over their anti-matter partners in the Universe. Subsequently, during post-doctoral periods at CERN and Dublin, I investigated specific modifications of standard cosmologies, in which the energy driving the expansion of the Universe is in part of an unusual type (so-called ``dark energy'').

Since about 1998 the underlying motivation in most of my research has been what is called the problem of ``large scale structure formation'' : the Universe today is observed to be highly inhomogeneous up to very large scales, with galaxies organised in a highly structured way - in walls, voids and filaments. The theoretical challenge is to understand fully these observations within a coherent theoretical framework. Modern cosmology provides such a framework, and even a very specific model (so called ``Lambda CDM'') which has had extraordinary success in explaining a whole host of observations in the last couple of years, most notably the fluctuations of the intensity of the ``primordial background'' of microwave radiation which have been measured in great detail by a series of experiments. However, the clear successes of the model concern the very large scales at which the Universe is close to homogeneous. Very many questions remain open concerning its capacity to explain the great wealth of data on the Universe at the smaller (but still very large !) scales at which it becomes strongly inhomogeneous.

My research aims to tackle fundamental, and quite general, open questions in this context. Just one example is the issue at the centre of a number of my publications in the last few years (and the subject of the thesis of my doctoral student B. Marcos): that of ``discreteness effects'' in huge numerical calculations of structure formation in cosmology. These calculations represent crudely the smooth distribution of the matter in these models by large self-gravitating clumps (called ``macro-particles''), and the question is how this artificial feature affects the precision of the results. This --- at first sight technical --- question is actually a very interesting and complex physical one, which involves understanding basic and unresolved issues about the clustering of many gravitating bodies, a problem which goes back to Newton.

What characterises my work and makes it unusual (and particularly original, I believe), is that it attempts to approach these problems from a different perspective to the usual one in cosmology : the perspective given by ``statistical physics'' which is the branch of physics dealing, in a general manner, with the understanding of systems of large numbers of constituant elements. My research is thus interdisciplinary, in the concrete sense that I collaborate with both cosmologists and statistical physicists. I do so because I am convinced that the methods of statistical physics, which are much more general than those of cosmology, can throw new light on the problems in cosmology, providing new tools and approaches.