|
|
|
I am a Research Scientist at the (ISC-PIF), focusing on the interface between
computer science and complex systems, in particular
computational biology and bio-inspired computing. I
studied physics at the École Normale Supérieure (ENS), Paris and
completed my PhD in 1991 in neural networks and cognitive science.
After postdocs in Germany and France, and a detour through the
software industry of the San Francisco Bay Area, , and
occupied various posts at the University of Nevada, Reno (US),
Ecole Polytechnique, Paris (France), University of Malaga (Spain),
Drexel University (US), and CNRS Paris, before becoming a Full
Professor at Manchester Metropolitan University (UK) in January
2016. I eventually quit in November 2019 for personal convenience,
and returned to ISC-PIF.
|
|
|
|
The Centre for Advanced Computational Science (CfACS), formerly headed by
Martyn Amos,
is an interdisciplinary hub leading work on fundamental and applied
computer science with a focus on high-impact interdisciplinary
research in healthcare, security and smart cities, taking an open
and collaborative approach to solution development. The CfACS hosts
groups in data science, machine intelligence, smart infrastructure,
and human-centered computing.
|
|
I conducted collaborative research in computational biology toward
original methodologies and tools for the multiscale image processing
and theoretical modeling of multicellular development. In parallel,
I pursued research in bio-inspired computing for the design of
decentralized, autonomous systems inspired by morphogenesis, with
applications to swarm robotics, synthetic biology and
socio-technical networks. I supervised PhD students, led and/or
co-authored grant proposals, articles in top-ranking journals and
conferences. I was co-director of the 30-staff CfACS and headed its
Complex Systems Group, fostering and supporting colleagues’ research
activities and publication plans. I also sat on several University,
Faculty and School committees, developed cross-campus, UK and
international collaborations, and continued (co-)organizing various
workshops and summer schools.
|
|
|
|
|
The BioEmergences Lab, founded by
Nadine Peyriéras
on the CNRS campus in Gif,
develops original methodologies and tools for the in vivo
multiscale and multimodal observation, quantification and
multilevel theoretical modeling of biological processes. Its
strategies are the basis for a predictive understanding of the
morphogenesis of living organisms in normal and pathological
conditions, opening the way to new kind of pharmacology and
toxicology screening schemes.
|
|
I pursued my research activities (see 15. above), also supervising
the development of software platforms, PhD students and postdocs,
and lecturing at the Erasmus Mundus Complex Systems Master’s of
Ecole Polytechnique.
|
|
|
|
|
The engineering program was established in 1896, soon after the
founding of CUA, and the School of Engineering in 1930. Research
activities and graduate offerings have increased at a steady rate.
Today, the school offers BSc, MSc and PhD degrees in several areas:
biomedical, civil, mechanical, electrical engineering & computer
science.
|
|
I taught six semesters covering four different classes in computer
science, math and engineering: Java programming (Sophomores),
control systems (Juniors), operating systems (Seniors), and
differential equations (Graduates). I also led initiatives on
research projects, collaborated with colleagues and supervised
students.
|
|
The School
of Bioengineering, Science and Health Systems (BIOMED) hosts
39 core faculty and 20 research groups (such as bionanotechnology,
integrated bioinformatics, optical brain imaging, and tissue
engineering). It has launched many research initiatives with
partner institutions and industry, merging physical sciences,
computational sciences, and technology with life sciences and
medicine toward a new era of health systems.
|
|
I was formally affiliated (non-employee appointment) to
collaborate with the faculty on submitting proposals that focused
on computational modeling and simulation of morphogenetic
engineering systems.
|
|
|
|
The Research Group in Biomimetics (GEB),
is headed by Francisco J. Vico at UMA's Department of Computer Science.
Its research topics in computational biology and artificial life
aim at explaining how complex patterns and novel behavior emerge
from living matter. GEB also has a track record of technology
transfer, having applied bio-inspired methodologies to a wide spectrum
of projects in partnership with companies.
|
|
I conducted collaborative research, publications, and co-supervision
of PhD students, focusing on Artificial Life.
|
|
|
|
|
The Complex
Systems Institute, Paris Ile-de-France (ISC-PIF),
founded by Paul Bourgine, a former director of
CREA, is a multidisciplinary research
center and network (GIS: "Groupement d'Intérêt Scientifique")
sponsored by the Paris Region Ile-de-France as part of its
research initiative (DIM: "Domaine d'Intérêt Majeur"). It is
co-financed by 15 top-rank French academic partners—graduate
schools, universities, and national institutions (Ecole
Polytechnique, Ecole Normale Supérieure, Université Pierre et
Marie Curie, CNRS, INRIA, CEA, Inserm, etc.). Its mission is
to build a community of research in complex systems (large
sets of locally interacting elements creating a collective
behavior), to study common "questions" (self-organization,
emergence, autonomy, adaptation, etc.) across many "objects"
(molecular, cellular, cognitive, social, economic,
technological, environmental).
|
|
Beside my research activities, I co-managed
during my first year, and contributed to, the EU
Embryomics and
BioEmergences projects
(FP6-NEST, 6 partners, €3.15M, 2005-2009),
launched by
Paul Bourgine
and Nadine Peyriéras, whose goal was the
spatiotemporal reconstruction of the full cell lineage tree
underlying biological development, by image processing,
visualization, and agent-based simulation. I also directed
a PhD thesis on this topic ( project, by Julien Delile).
|
I was also in charge of leading and managing the institute during
two full years, in collaboration with a Steering Committee ("Comité
directeur") of 15 external advisors, senior researchers and
professors. In particular, I led the yearly renewal of ISC's
mandate and funding. I reported twice a year to an Executive Board
("Conseil d'administration") composed of our supporting partners
about the institute's program, activities and budget use. The
capital budget ('investissement") and operating budget
("fonctionnement") totalled about €1M/year. Scientific
orientations were overseen by an external Scientific Council of
foreign scholars. Capital budget was used to create ISC "branches"
hosted by its different academic partners, i.e., build/renovate
and equip office space to become complex systems research labs.
Another part was invested in a large computing cluster (1600 cores)
dedicated to complex systems modeling and numerical simulations.
Operating budget supported ISC's activities, including scientific
events (conferences, workshops, seminars) and educational programs
(summer school, thematic institutes, Master's curriculum), and the
resident staff of 10 researchers, 3 engineers and 3 admins who
coordinated them. After two years, I decided to hand over this
day-to-day management responsibility in order to dedicate myself
again to research. In sum, it involved contributing to, or supervising:
- Liaison with main sponsor, the Paris Region: yearly financing agreements, collaborations with other regional centers
- Liaison with 15 co-sponsoring academic institutions: partner fees, bi-yearly meetings of the Board of Trustees
- Liaison and co-organization of events with the National Network (RNSC) and Lyon Institute (IXXI) of Complex Systems
- Liaison with the Complex Systems Society (CSS): org. of ECCS conferences, submission of grant projects to EU-FP7
- Agreements for the creation and management of complex systems "branch" labs hosted by our academic partners
- Purchase, installation, management of computing cluster (devt. of a simulation platform, creation of a user committee)
- Organization of events: summer schools, conferences, workshops, thematic institutes, seminars, etc.
- Organization of calls for projects (seed funding) and position searches (recruitment of postdocs, engineers, admins)
- Creation and coordination of a European Master's in Complex Systems Science (Erasmus Mundus II Program)
- Coordination of a team of resident researchers, postdocs, engineers, PhD students, MSc interns, admin staff
- Development of Web/Wiki site, specification of a membership policy, launch of public communication campaigns
- Development of partnerships with industry and public decision makers ("how complex systems science can help")
- Preparation and organization of meetings of the Steering Committee, the Board of Trustees, and the Scientific Council
- Administrative tasks: HR, accounting, operations, logistics, etc.
|
|
The value added by ISC-PIF to the previous state of complex
systems research in the Paris Region (Ile-de-France) is
considerable: the ISC-PIF has played a unifying role, opening
bridges between institutions, creating new structures and
networks, increasing the regional and national scientific
potential, and promoting exchanges between science and civil
(non-academic) society. Complex systems science represents a new
attempt to conduct on a large scale a multidisciplinary
scientific approach that is unprecedented and varied.
The Ile-de-France region is one of the great scientific regions
in Europe and in the world. It has a duty to be a major player
and, more importantly, a pioneer when a new field of science
emerges. No complex systems research structure existed there
prior to the launch of ISC-PIF by Paul Bourgine. Thus, before
federating a "complex systems community" by creating a network
for advanced studies, it was necessary to identify the
cross-cutting questions of complex systems (such as
"self-organization", "emergence", "robustness", "control",
"reconstruction of the dynamics", etc.) and the
researchers interested in these questions.
|
|
|
|
The UNR Brain Computation Laboratory
("Brain Lab") is an interdisciplinary research group studying
large-scale spiking neuronal models of the cortex. Its core
technology was the NeoCortical Simulator (NCS), a biologically
detailed software model running on a massively parallel 220-CPU
Beowulf cluster.
|
|
I was a Co-PI in the "Neuromorphic Mesocircuits" project
led by Philip H. Goodman (Lab Director & Professor,
School of Medicine; deceased in 2010). It constituted an original
attempt to design a modular brain architecture of spiking neural
networks that emulated robotic behavior learning. We modeled
pattern recognition and association by "lock-and-key" coherence
induction between dynamic cell assemblies.
I also further developed
the research on spatial categorization started at CREA, Paris
(emergence of symbolic language from visual scenes; see next)
and became actively involved in several other complex systems projects
(see above). Additionally, as a visiting faculty in the Department
of Computer Science and Engineering, I taught two to three classes
per semester, organized and co-managed student projects, and
assisted supervising MSc and PhD works.
|
|
|
|
|
|
|
|
The Centre de Recherche en
Epistémologie Appliquée
(CREA) is an interdisciplinary theoretical research lab
in cognitive and social sciences. Its activities range from
neuroscience to linguistics and economics, focusing on the
mathematical and computational modeling of complex,
self-organizing systems.
|
|
I worked with
Jean Petitot
(Professor & Director; also at EHESS—School for Advanced Studies
in Social Sciences, Paris) on dynamic models of semantics
based on cognitive linguistics (in contrast to logical models of
syntax based on generative grammar). We specifically examined spatial
categorization, i.e., how the mind is able to map an infinite
variety of visual scenes to only a few prepositions ('in', 'over',
'across', etc.). This study addressed central theoretical questions
such as the interface between physicalist and symbolic
representations and the existence of a "cognitive topology" in
perception (less metric than vector spaces, yet more metric than
topology). I created a graphical application to illustrate the
schematization pathways underlying classification of space,
and I collaborated to a book by Jean Petitot on these topics.
|
|
|
|
The Institute for Neural Computation (INI),
founded in 1990, is a research center in neural networks,
computer vision, neurobiological models, machine learning and
autonomous robotics. Its goal is to understand the organizational
principles of nervous systems and find new solutions to problems
of information processing in technical systems, such as real time
vision, face recognition, and medical engineering.
|
|
I worked under the supervision of
Christoph von der Malsburg (Professor &
Chair; also at the University of Southern California) on theoretical
aspects of pattern recognition, specifically the ability of the
visual system to segment and regroup image domains under the
influence of previously learned shapes. My focus was studying
networks of coupled oscillating units and their properties of
emergent collective behavior, such as phase-locking synchronization
or traveling waves of activity. I designed models showing that
shape extraction can arise from such networks and created network
simulator applications with high-end graphical user interfaces to
support these models. I also co-created and taught
two original seminars for graduate students in cognitive science,
on neuro-inspired learning and linguistic questions.
|
|
|
|
The Electronics Lab
at the School of Industrial Physics and Chemistry
(Ecole Supérieure de Physique et Chimie Industrielles,
ESPCI) is an engineering research lab in machine learning,
neural networks and signal processing, led by
Gérard Dreyfus, Professor & Director.
|
Under the direction of
Elie Bienenstock, CNRS (now Associate
Professor at Brown University, Providence), I elaborated a
criticism of the traditional activity-rate code in neural
models, which also advocated temporal correlations as the
basis of brain function (after von der Malsburg's theory, 1981).
We illustrated this question through three mathematical
and numerical studies:
- the bias/variance trade-off faced by machine learning,
like any nonparametric statistical estimation
- handwritten character recognition based on 2-D "elastic"
template matching (instead of pixel lists)
- synaptic self-organization in the cortex ("synfire chain"
growth) by activity/connectivity feedback
I co-designed the models, created visualization tools and carried
out numerical simulations for all three parts (see
).
|
|
iscpif.fr
