Robin Lamarche-Perrin

Printer-friendly versionPDF version

Contact Information

Laboratoire d'Informatique de Grenoble (UMR 5217), Magma Team
Office G120, Maison Jean Kuntzmann, 110 avenue de la Chimie
38041 Grenoble CEDEX 9, France

Email: Robin.Lamarche-Perrin <at> imag.fr
Phone: +33 (0)4 76 51 46 33
Mobile: +33 (0)6 26 01 32 91


I am now in a postdoctoral position at the Max-Planck-Institute of Leipzig.
Here are my new contact details:

Max-Planck-Institute for Mathematics in the Sciences
Research group Jürgen Jost, Office C10

Address:
Max-Planck-Institut
Inselstraße 22
04103 Leipzig, Germany

Email: Robin.Lamarche-Perrin <at> mis.mpg.de
Phone: +49 (0) 341 - 9959 - 631
Mobile: +49 (0) 179 - 1818 - 906


Here is my Ph.D. thesis, the defense video, the slides in French and in English.

Here is my curriculum vitæ (not updated since May 2013).

Research Topics

  • Theory: macroscopic observation of natural and artificial systems, abstraction process, concept of emergence, Artificial Intelligence, epistemology, philosophy of science.
  • Methods: information theory, complexity theory, information visualisation, statistical analysis, multi-scale analysis, data clustering.
  • Applications: complex systems, Multi-agent Systems, large-scale distributed systems, Geographic Information Systems, social systems and social sciences.

Research projects

  • GEOMEDIA, Observatoire des flux géomédiatiques internationaux, 2013-2015. Interdisciplinary project of the French National Research Agency (ANR-GUI-APP-04) bringing together geographers and media experts from the International College of Territorial Sciences and computer scientists from the Grenoble Informatics Laboratory. My role in this project consists in providing abstraction tools (WP3) for the spatial and temporal analysis of international relations through print media (WP4). See my video regarding this project (with english subtitles).
  • SONGS, Simulation Of Next Generation Systems, 2012-2015. Computer science project of the French National Research Agency (ANR-11-INFRA-13) interested in the simulation of large-scale distributed computing platforms. My role in this project consists in providing abstraction tools to tackle and visualise very large execution traces (WP7). In particular, it leads to the implementation of an aggregation unit within the open-source VIVA software for traces analysis.

Publications in Computer Science

National Peer-reviewed Journals
  • Robin Lamarche-Perrin, Lucas Mello Schnorr, Jean-Marc Vincent and Yves Demazeau. Agrégation de traces pour la visualisation de grands systèmes distribués. In Jean-Daniel Fekete and Pierre Dragicevic (eds.), Technique et Science Informatiques (TSI), 2014. (forthcoming)
International Peer-reviewed Conferences with Proceedings
  • Robin Lamarche-Perrin, Yves Demazeau and Jean-Marc Vincent. The Best-partitions Problem: How to Build Meaningful Aggregations. In Yi Pan and Vijay Raghavan (eds.), Proceedings of the 2013 IEEE/WIC/ACM International Conference on Intelligent Agent Technology (IAT'13), pages XX-XX. 2013. (forthcoming)
  • Timothée Giraud, Claude Grasland, Robin Lamarche-Perrin, Yves Demazeau and Jean-Marc Vincent. Identification of International Media Events by Spatial and Temporal Aggregation of RSS Flows of Newspapers. Application to the Case of the Syrian Civil War between May 2011 and December 2012. In Proceedings of the 18th European Colloquium on Theoretical and Quantitative Geography (ECTQG'13), p. 112-114, Dourdan, France, Sep. 2013. (proceedings)
  • Robin Lamarche-Perrin, Yves Demazeau and Jean-Marc Vincent. How to Build the Best Macroscopic Description of your Multi-agent System? In Yves Demazeau and Toru Ishida (eds.), Proceedings of the 11th International Conference on Practical Applications of Agents and Multi-Agent Systems (PAAMS'13), vol. 7879 of LNCS/LNAI, p. 157-169. Springer-Verlag Berlin, Heidelberg, 2013. (article, slides)
  • Robin Lamarche-Perrin, Yves Demazeau and Jean-Marc Vincent. Analysis of News through Spatial and Temporal Aggregation. In Yves Demazeau and Toru Ishida (eds.), Proceedings the 11th International Conference on Practical Applications of Agents and Multi-Agent Systems (PAAMS'13), vol. 7879 of LNCS/LNAI, p. 296-299. Springer-Verlag Berlin, Heidelberg, 2013. (article)
National Peer-reviewed Conferences with Proceedings
  • Robin Lamarche-Perrin, Yves Demazeau and Jean-Marc Vincent. Organisation, agrégation et visualisation d'informations médiatiques. In Proceedings of the 2011 International Conference of the International College of Territorial Sciences (CIST): “Fonder les sciences du territoire”, p. 240-246, Paris, France, Nov. 2011. (article, slides)
  • Robin Lamarche-Perrin, Yves Demazeau and Jean-Marc Vincent. Observation macroscopique et émergence dans les SMA de très grande taille. In Emmanuel Adam and Jean-Paul Sansonnet (eds.), Actes des 19es Journées Francophones des Systèmes Multi-Agents (JFSMA'11), p. 53-62, Valenciennes, France, Oct. 2011. (article, slides)
Oral Communications in National Peer-reviewed Conferences without Proceedings
  • Robin Lamarche-Perrin, Jean-Marc Vincent and Yves Demazeau. Informational Measures of Aggregation for Complex Systems Analysis. In Poster Session of the 2012 European Conference on Complex Systems (ECCS’12), Brussels, Belgium, Sep. 2012. (poster)
  • Robin Lamarche-Perrin, Yves Demazeau and Jean-Marc Vincent. Informational-Theoretic Measures of Aggregation for the Analysis of Complex Systems. In Journée des doctorants du Laboratoire d'Informatique de Grenoble (LIG), Grenoble, France, June 2012. (poster, slides)
  • Robin Lamarche-Perrin. Conceptualisation de l'émergence : dynamiques microscopiques et analyse macroscopique des SMA. Actes de la plateforme AFIA 2011 : Atelier Futur des Agents et des Multi-Agents (FUTURAMA'11), Chambéry, France, May 2011. (article, slides)
Other Publications
  • Robin Lamarche-Perrin. Analyse macroscopique des grands systèmes. Émergence épistémique et agrégation spatio-temporelle. Ph.D. thesis in Computer Science, supv. Yves Demazeau and Jean-Marc Vincent, Université de Grenoble, France, Oct. 2013. (thesisdefense video, slides in French, slides in English)
  • Robin Lamarche-Perrin, Yves Demazeau and Jean-Marc Vincent. Multi-resolution Representations of Media Information. In Video Competition of the 23rd International Joint Conference on Artificial Intelligence (IJCAI'13), Beijing, China, August 2013. (video, english subtitles)
  • Robin Lamarche-Perrin, Yves Demazeau and Jean-Marc Vincent. The Best-partitions Problem: How to Build Meaningful Agregations? Research report RR-LIG-044, Laboratoire d’Informatique de Grenoble, France, Feb. 2013. (report)
  • Robin Lamarche-Perrin, Yves Demazeau and Jean-Marc Vincent. How to Build the Best Macroscopic Description of your Multi-agent System? Application to News Analysis of International Relations. Research report RR-LIG-035, Laboratoire d’Informatique de Grenoble, France, Jan. 2013. (report)
  • Robin Lamarche-Perrin, Lucas Mello Schnorr, Jean-Marc Vincent and Yves Demazeau. Evaluating Trace Aggregation Through Entropy Measures for Optimal Performance Visualization of Large Distributed Systems. Research report RR-LIG-037, Laboratoire d’Informatique de Grenoble, France, Dec. 2012. (report)
  • Robin Lamarche-Perrin, Jean-Marc Vincent and Yves Demazeau. Informational Measures of Aggregation for Complex Systems Analysis. Research report RR-LIG-026, Laboratoire d’Informatique de Grenoble, France, May 2012. (report)
  • Robin Lamarche-Perrin, Yves Demazeau and Jean-Marc Vincent. Macroscopic Observation of Multiagent Systems. Research report RR-LIG-010, Laboratoire d’Informatique de Grenoble, France, March 2011. (report)
  • Robin Lamarche-Perrin. Observation macroscopique pour l’analyse de systèmes multi-agents de très grande taille. Master thesis in Computer Science, supv. Jean-Marc Vincent and Yves Demazeau, Université Joseph Fourier, Grenoble, France, June 2010. (thesis, slides)
  • Robin Lamarche-Perrin. Argumentation, contexte et alignement d’ontologies. Research work, supv. Jérôme Euzenat, Ensimag, INP Grenoble, France, June 2009. (report)
Other Oral Communications
  • Robin Lamarche-Perrin, Yves Demazeau and Jean-Marc Vincent. Agrégation spatiale et agrégation temporelle. Second Meeting of the GEOMEDIA Project (ANR-GUI-APP-04), Paris, France, 25 Feb. 2013. (slides)
  • Claude Grasland, Robin Lamarche-Perrin, Vega Partesotti and Marta Severo. La médiatisation des catastrophes : l'exemple des tremblements de terre et des tsunamis. Annual seminar of the International College of Territorial Sciences (CIST) : “Territoires en action”, Paris, France, 29-30 Nov. 2012. (slides)
  • Robin Lamarche-Perrin, Yves Demazeau and Jean-Marc Vincent. Mesurer, évaluer, agréger l'information médiatique. First Meeting of the GEOMEDIA Project (ANR-GUI-APP-04), Paris, France, 22-23 Oct. 2012. (slides)
  • Robin Lamarche-Perrin, Yves Demazeau and Jean-Marc Vincent. Data Aggregation for Complex Systems Analysis. Young Researchers Network Complex Meeting, Brussels, 2 Sept. 2012. (slides)
  • Davide Beraldo, Juan Galán Páez and Robin Lamarche-Perrin. Multi-dimensional Analysis of Media. Sixth Annual French Complex Systems Summer School (CSSS'12), Paris, France, 2-13 July 2012. (slides)
  • Robin Lamarche-Perrin, Yves Demazeau and Jean-Marc Vincent. Agrégation de données pour l'analyse de systèmes complexes. Annual Retreat of the LIG-MAGMA Team, Grenoble, 27 June 2012. (slides)
  • Robin Lamarche-Perrin, Jean-Marc Vincent and Yves Demazeau. Abstraction et observation macroscopique des systèmes. Seminar of the LIG-MESCAL Team : “Après-midi traces”, Grenoble, 20 March 2012. (slides)

Research Interests in Computer Science

My PhD began in October 2010 in the computer science laboratory of Grenoble named LIG (Laboratoire d'Informatique de Grenoble). My thesis is supervised by Yves Demazeau from the MAGMA team and by Jean-Marc Vincent from the MESCAL team. My work is funded by a grant attributed by the Ministry of Higher Education and Research to the UdG (Université de Grenoble).

This Ph.D. thesis is interested in the observation and the analysis of natural or artificial systems. When it comes to large-scale or very complex systems, the observation of microscopic dynamics may suffer technical and semantic limits: How to observe a million of distributed entities? How to interpret such a cluttered microscopic description? The concept of epistemological emergence inherited from philosophy, and more particularly from epistemology, offers an interesting theoretical framework to tackle this issue. Since microscopic analysis is not sufficient to explain complex phenomena, scientists need to define and use high-level abstractions to globally represent the system. Abstractions are thus interpreted as macroscopic observation patterns that reduce the system’s microscopic complexity.

In this thesis, we focus on a particular abstraction process, namely data aggregation. For the abstractions to be interpretable at scale, we propose to quantify and control the complexity reduction and the information losses induced by this aggregation process. Feedback techniques, using measures from information theory, guarantee that the observer is able to properly interpret the generated macroscopic descriptions. Moreover, to be meaningful, aggregation should fit with the semantic properties of the observed system. Exogenous constraints from the expert domain are exploited to lead the abstraction process. Thus, our aggregation technique allows to build multi-resolution descriptions of the system that both preserve its information content and its semantic properties.

The technical contribution of our approach is demonstrated with the analysis of large-scale parallel applications from High Performance Computing. The aggregation technique is successfully applied to the spatial aggregation of one million processes. It reveals anomalies at different levels without going through the complete visualisation of the microscopic level. The semantic contribution is evaluated within an interdisciplinary project (GEOMEDIA) interested in the understanding of international relations through print media analysis. Spatial and temporal aggregation allows to detect critical events at different scales, thus building strong abstractions to describe the local and global dynamics of the underlying social system. We also work on the observation of large-scale complex Multi-agent Systems. In this case, the aggregation process is itself an emergent process. The aggregation algorithms are distributed in space and time and directly provides macroscopic descriptions of the system, without observing in details all the agents.

EducatioN in computer science

2010 – present Ph.D. degree in Computer Science
Supv.: Y. Demazeau, J.-M. Vincent
Laboratoire d’Informatique de Grenoble
Université de Grenoble
2009 – 2010 Master's degree in Computer Science “Artificial Intelligence and the Web”
Supv.: J.-M. Vincent, Y. Demazeau
Laboratoire d’Informatique de Grenoble
Université Joseph Fourier
2007 – 2010 Engineering degree in Computer Science and Applied Mathematics

Ensimag
Grenoble INP

Publications in Philosophy

National Peer-reviewed Journals
  • Robin Lamarche-Perrin. Des collaborations possibles entre Intelligence Artificielle et philosophie de l’esprit. In Anna C. Zielinska (ed.), Repenser les rapports entre sciences et philosophie, collection "Recherches sur la philosophie et le langage", vol. 29, p. 47-65, Vrin, Paris, France, 2013. (sous-presse)
Oral Communications in National Peer-reviewed Conferences without Proceedings
  • Robin Lamarche-Perrin. Des collaborations possibles entre Intelligence Artificielle et philosophie de l’esprit. In 3e Colloque des doctorants et des jeunes chercheurs du groupe Philosophie, Langages & Cognition (PLC) : “Repenser les rapports entre science(s) et philosophie”, Grenoble, France, June 2011. (slides)
Other Publications
  • Robin Lamarche-Perrin. Des collaborations possibles entre philosophie et Intelligence Artificielle. Master thesis in Philosophy, supv. Denis Perrin, Université Pierre-Mendès-France, Grenoble, France, Sep. 2012. (thesis, slides)
  • Robin Lamarche-Perrin. Le Test de Turing pour évaluer les theories de l’esprit. Master thesis in Philosophy, supv. Max Kistler, Université Pierre-Mendès-France, Grenoble, France, Sep. 2010. (thesis, slides)

Research Interests in Philosophy

I recently obtained a Master degree in philosophy at the UPMF (Université Pierre-Mendès-France, Grenoble) under the supervision of Dennis Perrin from the PLC laboratory. I am there interested in the relationship between Artificial Intelligence (AI) and philosophy of mind as particular cases of collaborations between science and philosophy. Thereby, I work on particular interdisciplinary topics such as the concept of emergence and the concept of representation in AI.

Education in philosophy (in parallel)

2009 – 2012 Master's degree in Philosophy “Histoire de la philosophie et philosophies du langage”
Supv.: D. Perrin, M. Kistler
Université Pierre-Mendès-France
2007 – 2009 Bachelor's degree in Philosophy Université Pierre-Mendès-France