Inria - Activity report 2010


Published on

The 2010 annual report, covering the activity of the whole Institute, is now available in two formats: in the print version and in the on-line version that can be consulted online. It is an opportunity to look back over an eventful year and to share this document which is both important and at the same time enjoyable to read.

Published in: Technology, Education
1 Like
  • Be the first to comment

No Downloads
Total Views
On Slideshare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Inria - Activity report 2010

  1. 1. Annual report 2010
  2. 2. 02 _ In the words of the researchers08 _ Invent to serve09 _ Editorial12 _ A new organisation for a new way of working20 _ More structured and ambitious partnerships25 _ Transfer for innovation: Using research to tackle contemporary challenges30 _ the PLeasUre of InventInG40 _ InventInG In the fIeLd50 _ MasterMIndInG the fUtUre58 _ Debates70_ Staff, contributors and project-teams74 _ Project-teams active in 201080 _ Inria’s partnerships81 _ Organisation chart and councils
  3. 3. Inventors for the digital worldOur society relies increasingly on digital technologies to communicate,seek medical information, travel, or have fun. These often-invisibletechnologies simplify our tasks and enrich our daily lives, while alsodeveloping the economy.At the interface of computer science and mathematics, from pureresearch to technological development and to industrial transfer,researchers at Inria, a public research institute, are inventing tomorrow’sdigital technologies. Inria’s research is collaborative, which is evidenced bythe diversity of the talent comprising its research teams, as well as inthe many joint projects conducted with public and private researchentities in France and abroad.While competing with the leading international specialists in their field,Inria researchers and staff are also committed to sharing their knowledgewith the widest possible audience. InrIA AnnuAl repOrT 2010 /1
  4. 4. DIGITAL PRIVACY“The part that IT plays in our daily environmentis undeniably a source of progress… but onewhich could quickly turn into a nightmare ifwe do not take care to protect everyone’s privacy.”Claude Castelluccia, senior researchscientist, Planète project-team,Grenoble – Rhône-Alpes research centreThe Planète team’s research objective is toanalyse existing software and digital services,such as social networks and smartphones.In a world where everything is connected…and traceable, this work aims to develop newsolutions to protect users’ privacy, whilst stillmaintaining optimal service quality.2/ InrIA AnnuAl repOrT 2010
  5. 5. InTeRneT“Bringing a bit of semantics into a world of links willlead to better management of the data posted andviewed on the internet, and make it easier to search anduse it. The purpose of my work is to give internet usersthe means to control their presence on the Web.”Fabien Gandon, Edelweiss project-team,Sophia Antipolis – Méditerranée research centreThe edelweiss team’s research is devoted toreasoning in relation to the growing mass ofonline data. The researchers look at metadata,i.e. data that characterise other data in order tosort them and assess their pertinence.Having control over metadata means being able,in future, to have control over the Web. InrIA AnnuAl repOrT 2010 /3
  6. 6. enVIROnMenT-eneRGY“Mathematics lets you see the invisible. Our project involvesproviding a precise mapping of the subsoil, in order to optimisethe exploitation of fossil energy sources (petroleum, gas).We are also looking to model destructive geophysical phenomenato determine high-risk zones and thereby be able to protectlocal populations.”Hélène Barucq,Magique-3D project-teamleader, Bordeaux – Sud-Ouestresearch centre4/ InrIA AnnuAl repOrT 2010
  7. 7. The Magique-3D team works on seismicimaging and computing for seismologicalpurposes. It develops digital methods tocollect, process and aggregate large amountsof information. It builds complex models andmakes use of high-performance computingresources in order to apply them to geophysicalstudies. Watch InrIA AnnuAl repOrT 2010 /5
  8. 8. HeALTHCARe “Depression is a major public healthcare issue. Through my work, I seek to improve nerve stimulation treatment response. The solutions we’ve developed help to treat a large number of patients every day at Rennes hospital.” Pierre Hellier, research scientist in the Serpico project-team, Rennes — Bretagne-Atlantique research centre. It was in the Visages project-team that Pierre Hellier contributed to the concept of a neuronavigation system. The Visages team developed a neuronavigation device, a veritable “brain GPS,” that is used to guide medical and surgical procedures in real time in a simple, precise way. This image-processing control system is used in transcranial magnetic stimulation, a treatment for depression.6/ InrIA AnnuAl repOrT 2010
  9. 9. DISABILITY“Central nervous system injuries can cause paralysisin certain limbs even if the muscles are intact. In mywork, I aim to assist or restore functional movementin paralysed limbs using their residual motor abilities,particularly in cases of hemiplegia and paraplegia.”Christine Azevedo Coste,researcher in the Demar project-team,Sophia Antipolis – Méditerranée research centre The Demar team strives to improve electrical stimulation methods, i.e. the skillfully calculated application of electrical currents that trigger coordinated muscle contractions in the paralysed limbs. To obtain functional movement (walking, posture, prehension), it is necessary to control the activities of all the muscles involved: those of healthy as well as disabled limbs. InrIA AnnuAl repOrT 2010 /7
  10. 10. 8/ InrIA AnnuAl repOrT 2010
  11. 11. Inria is readyfor this decadeI n a digital age, Inria is ready, now more than ever, to handle all of its missions: facing new scientific challenges, developing relations with the economic and industrial world, participating in major societal projectsand raising public awareness of digital sciences. Inria has proven its capacityto evolve, demonstrating its scientific agility and solid foundations, in a con-stantly changing French research environment. Michel Cosnard, Chairmanand CEO of Inria, and Antoine Petit, DeputyManaging Director, share their views.In a period marked by continuousupheavals, for society in generaland in the world of research, how canthe Institute ensure and maintainits vibrancy?Michel Cosnard: Since the world and societyare becoming more digital, the need for re-search and innovation in our scientific fields isconstantly growing. New research themes areemerging, linked to our traditional skills, atthe crossroads of informatics and mathematics.They are more systematically linked to societalchallenges like the environment and health,the internet of the future (sensor networks,social networks…), security and reliability ofsoftware programmes or privacy issues. In this
  12. 12. particularly stimulating context and in order to respond to new problems, Inria researchers are launching new projects, looking for new partners and forming new teams. Antoine Petit: For all these challenges, competition is global. Indeed, beyond their cognitive aspects, digital technologies are considered, and rightly so, as a key factor in economic and social development, thanks to their capacity for innovation and disseminating technology. Our ambition is clearly to continue working within the exclusive circle of leading research institutes that are an international reference in digital sciences. How has Inria been able to grow and control its development? M. C.: Above all, there are the fundamentals that characterise our work. The first is the nature of our research teams: our teams are small and work on projects with strong societal or economic implications. The second concerns our eight research centres, which all have, working alongside researchers, departments dedicated to development and support for research. This is an important asset. Finally, the national character of our institute translates into the definition of strategic national schemes, implemented in a transdisciplinary manner by our scientific and functional departments. Our organisation does not have a rigid hierarchy. It is capable of rapidly evolving to tackle new subjects in different fields, thus ensuring great agility. This is true for our research teams, but also for our different business lines, which regularly demonstrate their ability to adapt. A. P.: Inria’s flexible organisation allows it to respond effectively to a certain number of current scientific and organisational challenges. Thanks to our responsive project-teams, and their appeal, we can focus our efforts on new projects and attract talents from around the world. For example, Inria mobi- lised its resources in the context of the Future Investments programme. The10/ InrIA AnnuAl repOrT 2010
  13. 13. x2 In ten years, the institutepurpose for Inria was to confirm its role as a national has doubled in size and budget.player in transfer and innovation - as can be witnessedin our Institut Carnot certification - to cultivate newpartnerships in life sciences, the environment and medicine, and to consolidateits influence in a digital economy.What are some major projects in the future that will helpInria confirm its position and ensure its missions?A. P.: Inria has a long tradition of working with the industrial and economicworld, for which it is a privileged partner in digital technologies. We aregoing to pursue this policy, by focusing our efforts on SME. Moreover, weintend to develop our relations with the public at large, in order to answerquestions our fellow citizens have about digital technologies and their rolein society. As a national institute, Inria also intends to continue being akey player in European research policies and contribute to constructing theEuropean research area.M. C.: In ten years the Institute has doubled in terms of size and budget. Nowour goal is to consolidate our organisation, assert our national presence, mul-tiply our partnerships with French and European universities and reinforceour complementary role with the CNRS. Therefore, our research centres,which are strongly embedded in their regional environments, should playa major role in elaborating and building territorial strategies for excellencein digital sciences, in partnership with all the players concerned. Thus, theywill contribute to creating the strong and highly visible regional clustersFrance needs. To sum up and conclude, today we are equipped to face thescientific, social and economic challenges of the next decade, in which digitaltechnologies will play a major role. InrIA AnnuAl repOrT 2010 /11
  14. 14. A new organisationfor a new way of working M ore modern management, clearer relations with other institutes, focusing of investments on fledgling centres, rolling out of the Allistene alliance, an increase in its own resources: Inria has con- solidated its organisation and ensured the longevity of its way of working in a changing research environment. “2010 has been devoted to a lot of work in order to prepare for the future,” con- cludes Hervé Mathieu, Chief Executive Officer for Resources and Service Administration. “We had to adapt to new rules and a new development model promoted by the State. A lot of projects were conducted, including some that will essentially be completed in 2011.” The goal is to prepare the institute for new challenges: reorganisation of the world of research, new scientific challenges related to societal issues, adaptation to changes in how organisations are funded and the development of new tools. 45 A complete reorganisation of management methods and procedures Projects dedicated to modernising management were launched in 2009 and led to certification of the institute’s accounts in 2010. framework agreements signed This regulatory obligation required a great deal of work and a with partners in higher education and research since 2009. complete review of financial procedures, which, after examina- tion by the statutory auditor, only gave rise to a small number of reservations. At the same time, reorganisation of IT means and a necessary update of the institute’s internal information system progressed significantly and will continue in 2011. This12/ InrIA AnnuAl repOrT 2010
  15. 15. The Saclay cenTre: A key sITe FOr InrIA Nozha Boujemaa, Director of the Inria Saclay Ile-de-Francephase, devoted to harmonising practices research centreand sharing data, represents an inestima-ble move forward in managing the insti-tute. It will make managing research pro- “The efforts deployed by the Inria To maintain skills in our scientificjects easier, for example in the context of for the Saclay centre correspond disciplines, it is vital that thisEuropean contracts or the ANR. to the challenges being tackled increase in the number of here, in an extremely rich researchers be accompanied bypositive growth, focused on scientific environment, at the an increase in human andnew centres heart of the ‘French Silicon Valley’ material resources to organiseThe new model for organising research promoted by the government. the centre and its departmentsthat is taking shape has been accompa- The centre is highly involved in (legal support, exploitation andnied by a stabilisation in State funding: this project since 25 of its transfer of results, etc.) thus“Our growth is now supported by our own re- 28 teams are formed jointly with guaranteeing the performancesources,” points out Hervé Mathieu. “They other partners on the site of of research teams.”have been highly significant this year, with an Saclay.overall increase of 47% and growth of 28% forrevenues linked to research contracts. This hasallowed an increase of our means, in constanteuros, of more than 11% in 2010.” Therefore, despite a decline in the numberof openings for civil servants, recruitment levels have remained high. Invest-ment efforts have been concentrated on three recent centres in Bordeaux,Lille and Saclay, whose basic means have not been completely establishedand who have pressing needs in terms of infrastructure.A successful transition into the futureInria has aligned itself with the new national strategy in research and inno-vation (SNRI) adopted in July 2009 and which is organised around universi-ties. In particular, a framework agreement was signed in December 2009with the Conference of University Chancellors (CPU). “In 2010, we set up,within the context of the Allistene alliance, a common environment for work-ing with universities and schools as well as the CNRS (French National Centrefor Scientific Research) and the CEA (French Atomic Energy Commission) in InrIA AnnuAl repOrT 2010 /13
  16. 16. computational sciences and technologies,” says Hervé Mathieu. Coordination with the CNRS led, in 2011, to the signature of a framework agreement with the latter, clarifying, in particular, the organisation of partnerships set up by both institutions with universities, either within joint research units (UMR) or joint project-teams (EPC). The Allistene alliance should help identify common scientific and techno- logical priorities among its members in order to reinforce partnerships and create links with businesses. “The alliance’s structure has been defined this year, with the creation of a coordination committee chaired by Michel Cosnard, five in- terdisciplinary task forces and six policy groups that are now up and running,” explains Claude Kirchner, Executive Officer for Research and Technology Transfer for Innovation. “An agreement has also been signed with the ANR in order to specify how the two entities are going to establish a joint programme, taking into account the major challenges in the field. Furthermore, the Allistene coordination committee has voted in favour of setting up an ethics committee in computational sciences.” Finally, systems to support the creation of new firms have been reorgan- ised to increase efficiency and integrate more easily the common strategy of actors within the alliance. Thus, Inria-Transfert has been replaced by three firms: Inria Participation (a 100% subsidiary of Inria), which covers all Inria interests in firms, IT-Translation (IT2), which supports entrepreneurs, and IT2I, a venture capital fund set up in partnership with a state-owned bank, the CDC (Caisse des Dépôts et de Consignation), which will inject funds into start-ups. Organising science in order to be more efficient and responsive Consolidated in 2010, the new organisation of Inria’s research activities is de- signed to maintain its efficiency and responsiveness, even though its research teams have doubled over the last decade. Research activities are organised ac- cording to five major fields, each led by a Deputy Scientific Director (DSA). This system allows them to work closely with researchers and coordinate their activities more effectively, notably in terms of collective responses to14/ InrIA AnnuAl repOrT 2010
  17. 17. “helping emerging communiTieS geT OrgAnIsed” Gilles Dowek, Deputy Scientific Director, “Algorithms, Programming, Software and Architecturescalls for projects. “The DSA provides scientificleadership in his field and acts as coordinatorwho is very close to both teams and research- how do you envision the field today?ers,” emphasises Pascal Guitton, Director of Visiting the 43 teams in my field allowed me to understand howresearch department. “They must also con- quickly information technology is evolving. I was struck by the manytribute to implementing Inria’s national scien- revolutions that are unfolding simultaneously and in which Inria is playingtific policy and help shape collective visions in an important role. I also observed that communities with members fromorder to synchronise the work that needs to be different worlds are forming around a common interest, like security,done over the next five years. Finally, they must for example. My job is to identify these emerging groups and helphelp identify new research themes that the in- them get organised.stitute should get involved in.” Which project seems the most vital to you? Contributing to public debate and clarifying issues is one of Inria’sIn this context, “exploratory” incentive ac- missions, but the institute does not know how to respond in a concise waytions allow the institute to consider new to very general questions, for example, systems security or protectinghypotheses that break with conventional privacy. My role is to promote the emergence of intelligible discoursethinking, before creating a team on a topic, about our activities, whether directed to the public at large or ourif it indeed turns out to be promising. These supervisory ministries.initiatives are scientifically risky and there-fore limited in number. On the other hand,the number of large-scale initiatives, whichinvolve Inria teams and external partners over a 4-year period in the studyof complex issues, are destined to increase. “Indeed, these initiatives tackle verydifficult problems that a single team could not easily solve and therefore necessarilyinvolve several players,” explains Pascal Guitton. “Inria wants to engage in moreof these projects because they represent current scientific challenges in our fieldsof research and often involve major stakes for society such as health, energy or theenvironment.” Two large-scale initiatives have been created this year, raisingthe total number to seven.Major tools to organise software developmentAn essential step for innovative research is also the possibility of experiment-ing with algorithms and software on technological platforms. Providing InrIA AnnuAl repOrT 2010 /15
  18. 18. invaluable support for research teams, experimentation and development departments have also been deployed and organised in 2010. Inria has also invested, with institutional partners and local authorities, in high-tech facili- ties. In Lille, the institute inaugurated the Inria Euratechnologies platform, a place for exchanges and collaborative work open to Inria researchers and part- ners. It is home, in particular, to the Lille node of a network experimentation plat- form (Senslab) deployed on sites in Lille, Grenoble, Rennes and Strasbourg. In Nan- cy, the high-security computing laboratory has opened its doors. It is used to conduct experiments by Inria and its partners in network security, exchanges and telecom- munications equipment. Finally, a virtual reality room has been inaugurated at the Sophia Antipolis-Méditerranée Research Centre (see following p. 18-19). It will be open to regional or European academic A HOuse AdApTed To The diSabled and industrial partners. Promoting home support and independence of the elderly or disabled is the objective of a large-scale initiative called PAL (Personally Assisted Living). This scientific and technical challenge involves a wide range of skills: 9 teams from 4 of the institute’s centres and a host of partners, including the CSTB (Centre Scientifique et Technique du Bâtiment) and the nice CHU (teaching hospital). They are organised around a dedicated infrastructure where they can combine and experiment with innovations in robotics, sensors and cognitive methods to prevent falls, detect signs of malnutrition, improve mobility or preserve social ties.16/ InrIA AnnuAl repOrT 2010
  19. 19. WatchHIGH-SECURITY LABORATORY (LHS),Nancy – Grand Est centre. Computing andtelescope clusters for collecting and studyingIT threats. Wadie Guizani, engineer inthe Carte team. InrIA AnnuAl repOrT 2010 /17 InrIA AnnuAl repOrT 2010
  20. 20. Watch IMMERSIVE ENVIRONMENT FOR TREATING DOG PHOBIA The immersive cube consists of three vertical screens and a horizontal screen. The purpose of the system is to obtaina real immersion sensation for a single user. The cube is one of the components of the Gouraud-Phong room, an immersive space of variable dimensions at Inria Sophia Antipolis – Méditerranée. 18/ InrIA AnnuAl repOrT 2010
  21. 21. InrIA AnnuAl repOrT 2010 /19
  22. 22. More structuredand ambitious partnerships n ow more than ever, Inria’s work is part of a national project open to Europe and the world. Defining means of cooperation and agreeing on common research themes is essential for creating partnerships more easily and achieving a higher level of performance. Work on ambitious projects that the institute started in 2009 is starting to produce results. “Preparing for the future concerns all of our relations with academic and indus- trial partners,” underlines Claude Kirchner, Executive Officer for Innovation Research and Technology Transfer. “These relations are destined to grow stronger. They should be more ambitious and more thoroughly developed so that we can control transfer better and thus contribute to creating wealth. They must be organised in order to achieve greater performance and legibility.” This is the challenge Inria intends to take up by reinforcing the high-quality relations it has cultivated with research teams and industrialists. This 40% approach is deployed on national, European and interna- tional levels. A partnership for the environment In 2010, funding through bilateral industrial agreements and sustainable development increased by 40%. On 8 September 2010, Inria and the Cemagref signed a partnership agreement reinforcing their collaborative work on environmental issues and current challenges raised by global climate change and sustainable development. In fact, concerning the environment, as in many other disciplines20/ InrIA AnnuAl repOrT 2010
  23. 23. MOre sTrucTured And AMbITIOusparTnerShipS WiTh induSTry from a high level of expertise on and concern parallel programming important questions. It gains a long environments, optimisation of energy term vision that allows it to identify consumption or the resilience and challenges or potentially interesting fault-tolerance of tomorrow’s subjects for its business in the future.” computers. The second agreement, The first of these strategic partnerships with the AnDRA (French national was signed with Alcatel-Lucent and has radioactive waste management led to the creation of an open laboratory agency), involves six Inria teams that will celebrate its 3rd anniversary in in research on modeling and simulation 2010. “Around a dozen patent of physical and chemical processes To this end, Inria set up, in 2008, applications have been submitted that affect buried radioactive waste strategic bilateral agreements that jointly, based on three major research throughout its existence. For the contribute to organising the institute’s initiatives undertaken in 2008, and AnDRA, this partnership contributes, skills around major industrial projects with third parties have been among other things, to optimising the challenges. These partnerships are launched, notably within the 7th use of computation codes, which are based on ambitious common research Framework Programme for Research increasingly cumbersome, and programmes, for periods of 4 to 5 years, and Development,” emphasises Olivier managing ever-increasing amounts focused on themes that are important Trébucq. “It’s a wonderful illustration of of data. for industry such as simulation and the high level of skill and emulation that The third agreement, signed with eDF, high-performance computing for eDF, exists within the laboratory!” A real defines the terms of a partnership self-organised networks for Alcatel or community of expertise has been concerning high-performance security for Microsoft. A framework established around key issues for the computing and simulations for energy. agreement defines the conditions of internet of the future. 12 Inria project-teams work with 6 R&D this work in terms of intellectual departments at eDF to develop tools property rights, management or New agreements around strong for visualising large volumes of data publication issues. “The objective is to socio-economic challenges or programming models for hybrid promote quality partnerships that allow Three new framework agreements were multi-core architectures. Inria researchers to confirm their ideas finalised in 2010, raising the total Three start-ups created thanks to Inria and software in real life application number of strategic partnerships signed research (Distene, Sysfera and Caps scenarios provided by the industrialist,” to 9. The first one, signed with Bull, entreprise) are also involved in the explains Olivier Trébucq, Head of led to the selection of five projects on design of these research initiatives. Strategic Partnerships and Key themes defined with the firm. These Accounts. “As for the firm, it benefits projects mobilise a dozen Inria teams InrIA AnnuAl repOrT 2010 /21
  24. 24. involving complex problems, researchers need to rely on computational sciences. Modelling and simulation, observation and detection techniques, managing huge amounts of data, etc. are at the heart of Inria’s work and its commitment to serving other disciplines. creating a european network to reinforce innovation Today Inria is deeply involved in the major European alliance EIT ICT Labs and coordinates French initiatives. This Knowledge and Innovation Community (KIC) must reinforce synergies between research, training and innovation to support European industry in services and applications linked to the information society. This is a crucial sector for European economies. 2010 was devoted to organising this highly ambitious pro- gramme and its deployment on 5 sites (Berlin, Eindhoven, Helsinki, Paris and Stockholm) involving 21 main partners, including 8 industrialists, 6 research organisations and 7 leading universities, as well as innovation centres and competitiveness clusters. Since September 2010, the community has had a CEO, Willem Jonker, and a “co-located” system of governance involving the 5 centres. “In 2010 we created tools to encourage exploitation and transfer of research results,” explains Bruno Le Dantec, head of the French “node”. A European entrepreneur’s club now provides a 8 point of access for national firms interested in doing busi- ness in other member countries. At the same time, the Technology transfer program will facilitate the transfer of In 2010, eight of the highly prized research results to industrialists and SME on a European erc (european research council) scale. “This tool will greatly improve use and dissemination grants were awarded to of research results,” according to Bruno Le Dantec. “The researchers on Inria project-teams. European project Contrail in the field of cloud computing has benefitted from support that has allowed it to confirm its results on European test benches, before disseminating them widely among partners and European students through workshops Watch and a summer school session.”22/ InrIA AnnuAl repOrT 2010
  25. 25. a Second TrAnsnATIOnAl TeAMresearch and trainingthroughout europeAt the same time, research activities based on the Inria has created the first joint project-team with a europeanthemes of EIT ICT Labs have been set up. “The university and its second transnational team since the one itFrench part coordinates the theme Tomorrow’s digi- created in 2008 with a Dutch organisation, the CWI. Thistal city,” indicates Bruno Le Dantec. “It involves project-team, called Focus, is based at the University of Bologna inimagining services that will make life easier for urban Italy and offers a perfect example of the institute’s europeanresidents and integrating work conducted on other strategy. It is led by Davide Sangiorgi, an internationally renownedthemes (Health and wellness, Energy efficiency…). researcher in modelling mathematical calculations in the study ofWe have obtained permission from the mayors of distributed systems. This is a fundamental subject, but one thatmajor cities covered by the 5 nodes to carry out offers a response to a concrete need: mastering computer systemsexperiments in situ.” integrated in everyday activities and objects and on large-scaleConcerning training, 7 European Masters have networks.been created and will be taught in partneruniversities, in English, starting in 2012, thusencouraging the mobility of both students and professors. The French par-ticipate in three Masters: internet technology and architectures, Distributedsystems and services and Human computer interaction and design.consolidating partnerships with north AmericaOrganisation and legibility are also the watchwords of Inria’s policy con-cerning international relations. Created in 2009, the JLPC (Joint Laboratoryon Petascale Computing), a joint laboratory with the National center forSupercomputing Application at the University of Illinois (USA) has led tohighly productive collaborative projects.This has allowed French researchers to take part in a major American pro-ject, the Blue Waters Petaflop Computer, and to contribute to the design ofsoftware for the optimal operation of this supercomputer. 10 articles and5 software programmes have already been produced. Thanks to this high-quality contribution, the JLPC now coordinates an international researchprogramme on climate simulations based on high-performance computing, InrIA AnnuAl repOrT 2010 /23
  26. 26. initiated by the G8 and involving partners from the United States, Canada, Germany, Japan, Spain and France. Other collaborative work with American researchers should be more visible. “Around thirty of our teams have created ties with researchers at Berkeley and Stanford, which are among the leading universities in Inria’s fields of expertise,” notes Hélène Kirchner, Director of International Relations. “We would like to reinforce these relations and highlight them more.” This goal was formalised in 2010 with the signature of an agreement with the Citris (Center for Information Technology Research in the Interest of Society) in order to organise existing collaborative work within a common project called Inria@Silicon Valley. consolidating federal and national agreements Another project designed to formalise scattered SupporTing dynamic collaborative projects has started in Brazil. Inria AFrIcAn reseArcH has had long-standing relations with Brazilian researchers and co-funds exchanges between teams. The CARI, the African symposium on “Inria would like to reinforce, organise and consolidate computer science research, organised relations with different Brazilian states in order to assert its 10th edition this year. This event, its presence in this vibrant and promising country,” initiated by Inria and the United nations underlines Hélène Kirchner. A cooperation agree- University in 1992, has become ment signed with 11 federal research agencies and a reference for African and their coordination within a national confederation French-speaking researchers over the has enabled the launch, in 2010, of a joint call for years. Today, it is ready to move up to projects on behalf of all these states. Sixteen Franco- a larger scale and include other Brazilian projects have already been submitted. countries, including english-speaking ones. This could be the next step for the upcoming CARI in Algeria in 2012.24/ InrIA AnnuAl repOrT 2010
  27. 27. Transfer for innovationusing research to tacklecontemporary challenges s ince its creation, the mission of Inria has been to ensure the trans- fer of knowledge and technologies developed by its teams to indus- try. Its purpose? Making sure its work in R&D is transformed into products and services, contributing to the creation of economic value. Currently SMEs are the institute’s privileged partners in this joint effort to promote innovation. Sharing expertise, knowledge and new technologies with businesses is one of the missions the State has assigned the institute. This means, on the one hand, transferring knowledge through strategic partnerships with large industrial corporations, as described above. More broadly speaking, this also means transforming technologies developed through research into products and services launched on the market. This requires adaptations. To this end, Inria has set up a specific offer for SMEs in order to create new partnerships. In 2010 the institute also created an internal programme to support scientists conducting technology transfer projects. “To deliberately speed-up and energise this transfer we joined forces in 2010 with OSEO, a French organisation that supports innovative SMEs,” underlines David Monteau, Deputy Director of Transfer and Innovation. “The purpose of this association is to offer SMEs greater visibility concerning public research in our field. This also allows us to identify innovative sectors with strong potential for growth and levers for intensifying transfer of tech- nologies in these sectors.” InrIA AnnuAl repOrT 2010 /25
  28. 28. lyATIss, inria’S 100th STarT-up Pascale Vicat-Blanc, Research director at Inria speeding up transfers to sMesResearch director at the Inria, currently on a leave of absence, The OSEO-Inria partnership provides SMEs withfor ten years Pascale Vicat-Blanc led teams specialised in support a better understanding of the skills and technolo-networks for internet applications, grids or clouds, which are very gies developed by public research facilities and thedemanding in terms of speed and delivery. “As early as 2005, I had opportunities for development they offer. In 2010the idea of a new model for the internet, a model that is called Inria has multiplied opportunities for encounters‘Cloud 2.0’ today,” she confides. “I started several projects in France between researchers and business. Four one-dayand internationally and then, in 2009, I decided to formalise the national “Inria-Industry Meetings” were organiseddeployment of this approach, with the help of a small group of doctoral on specific themes (aeronautics, e-health, sustain-students who believed in the concept. With one of them, who won able cities), combining technological demonstra-the Marconi Young Scholar prize, I created LYaTiss.” tions and forward-looking workshops in order toThis start-up was the first to propose a platform focused on the highlight the expectations of both parties. Other,network, enabling the deployment and optimisation of dynamic more numerous and targeted, encounters arecomputing and communication infrastructures, and maximal organised in research centres, often leading toperformance and agility of externalised applications. “Internet users promising collaborative projects. Above all theoften evolve faster than research and put pressure on firms,” OSEO-Inria partnership has allowed two large-she explains. “ Thanks to our software, firms can effectively adapt scale initiatives in favour of innovation to be settheir online service offer and remain competitive.” up. The first, the Mobile Service Initiative, brings together a host of players in telephony and mobile services. The second, the HPC-SME Initiative, organ- ised with the GENCI (Grand Equipement National de Calcul Intensif), facilitates access to high-performance computing for SMEs, thanks to an expertise and support programme. economic models for disruptive technologies OSEO contributes its expertise to a programme designed to monitor transfer initiatives. “In general we do not know, a priori, the right path for a technology to enter the market. The programme helps us provide support for researchers who are involved in this process and to help them financially,” explains Bruno Sportisse, Director of Transfer and Innovation. Within this programme, professionals 26/ InrIA AnnuAl repOrT 2010
  29. 29. help researchers design their transfer project by takinginto consideration, in addition to technological aspects,economic and managerial concerns. The programme also “Transfer is a long and complexrelies on a committee of external experts that studies the process: the production cycle forprojects and makes recommendations. “Transfer is a long a research prototype can takeand complex process: the production cycle for a research proto- anywhere from a few months totype can take anywhere from a few months to several decades.” several decades.”Situations can therefore vary a great deal. The committeeof experts indicates the most relevant means of transfer (partnership, start-upor direct transfer) and provides information on the economic environment inwhich the project promoters will be working. Their common goal: making surenew products and services meet societal demands and that the correspondingeconomic models are sustainable and totally independent from the Inria. “Toachieve this, adds Luc Grateau, we often need to suggest rather radical changes in posi-tion… encouraging the project promoters to abandon a strictly technological vision offunctions in favour of a vision conditioned by demand, uses or regulations.” In 2010,39 projects entered the programme and 11 were transformed into start-ups and2 became I-Labs, which are light-weight laboratory organisations associatingSMEs and Inria project-teams.Opening and sharing expertise in transfersFinally, the programme for monitoring transfer initiatives can detect work thatshould be exploited and decide to support it. This is the case for the work ofthe Aviz team at Saclay on interactive visualisation of data. “Our work involves The first issue of Connect,technologies that French engineers do not master,” says Jean-Daniel Fekete, head the magazine for SMEs seekingof the Aviz team. “In order for these to be transferred, they must be a certain level to innovate through the digital sciences, came out in 2010.of industrial quality upstream. This is a specific model Inria decided to support, byallowing the Aviz to recruit an engineer for one year. His assignment will be to rewriteour most popular systems so that they reach this level of industrial quality.” Entirelydedicated to digital sciences, today the institute has recognised expertise inthe transfer of software technologies. In order to propose this expertise and Watch InrIA AnnuAl repOrT 2010 /27
  30. 30. open its systems to university partners, who are often generalists, the institute signed agreements with a dozen university clusters in 2010. On a European level, Inria has brought together research institutions recognised for the quality of their transfer activities in the software sector and currently supervises the EIT ICT-Labs Technology Transfer Program. The dual advantage of open source software “The open source software programmes we provide are all, first and foremost, objects of research,” underlines Stéphane Ubeda, Director of Technological Development. “They illustrate the nature of our work and demonstrate our know-how. To prove their value, we make them available to the scientific community and our industrial partners.” Open source software offers several advantages. “Making the soft- ware’s source code available forces us to achieve a certain level of quality,” indicates Stéphane Ubeda. It is above all the “open source community” that enriches the software, through a constant improvement process. “We help maintain this community, since its members use, adapt and develop open source software, adding value for everyone,” observes Stéphane Ubeda “Research is a world based on shar- ing: the goal is to multiply our efforts so the code remains effective.” “The community created around open source software inherently generates transfers of technology,” adds Patrick Moreau, who is in charge of software assets. This is one of the reasons behind the creation of the IRILL (Research and Innovation on Free Software): demonstrating that transfer can take place via open source research and development. “Half of the software we disseminate is open source: afterwards, we need to be concerned with what becomes of this software. We need to make sure it survives outside our walls and that users and publishers appropri- ate the work produced by our researchers,” explains Patrick Moreau. From this perspective, the IRILL can serve as a catalyst.Irill, the Initiative pour la recherche etl’innovation sur le logiciel libre (Research andInnovation on Free Software), was createdin partnership with Universities of Paris 6and 7 in October 2010. Watch28/ InrIA AnnuAl repOrT 2010
  31. 31. WatchSEEING, UNDERSTANDING, AND ACTINGAt a time when masses of information are increasingat an exponential rate, visual analytics combinesanalysis methods with interactive visualisationmethods to give the user initiative and control overthe analyses in view of results that have already beencalculated and visualised.
  32. 32. 30/ InrIa annual report 2010
  33. 33. every day, Inria researchers use their creativity,knowledge, intellectual curiosity and specificknow-how to invent the digital technologiesof tomorrow. theirs is a profession that is madeup of small joys and great discoveries, but alsofrequent self-inspection. presenting one’s firstresearch results with enthusiasm, imparting one’spassion by talking at conferences, facilitatingexchange in the scope of an international project…eight Inria researchers share memorablemoments from their lives as scientists. InrIa annual report 2010 /31
  34. 34. presenting your research is an unforgettableexperience“Explaining the result of your work before an audience of research-lecturers puts you in a thrilling situation. I experienced it during theannual congress at which the Gilles Kahn thesis prize is awarded toyoung researchers in computer science. It’s a real privilege to be ableto talk about a subject that you’ve worked on, and to share it with anentire conference audience. But you have to find the right words topopularise the topic, keep their attention, and highlight the potential ofyour research. Because it’s not only a question of convincing, but alsoof raising interest in order to establish future collaboration. The actualfact of having to reformulate is in itself very motivating: it’s an exercisethat forces you to think and look at the results of your research in anew way. An opportunity not to be missed to build your knowledge!”Xavier allamigeon software validation as a member he has developed and made publicA graduate from the École of the SE/IS research team at EADS a free-licensed tool. Recruited bypolytechnique, Xavier Allamigeon Innovation Works and the MeASI Inria the same year, he joinedreceived the Specif - Gilles Kahn laboratory of the French Atomic the Maxplus project-team whereaward in 2010 for his work on Energy Commission. As a result, he is pursuing his work.
  35. 35. I chose to apply my passion to research“I’ve always loved putting my brain to work on problem-solvingtasks. This character trait is the reason why I was attracted to appliedmathematics. It was while I was doing my Master’s degree and theninternship at Air Liquide R&D that I discovered operational research.My hobby could actually be put to professional use! I then dida Google search for words that have an important meaning forme – optimisation, probabilities, statistics, etc. – and found a PhDoffer! Rather than taking a job after my Master’s degree, I decidedto plunge into the world of research and work on combinatorialoptimisation, which consists of finding the optimal solution for aspecific problem. A lifetime of problems waiting to be solved – whatmore could I ask!”marie-ÉlÉonore marmion computer science, in the field of resolution of combinatorialAfter obtaining a Master’s combinatorial optimisation. In optimisation problems. She hasdegree in applied mathematics, 2008 she joined Inria’s Dolphin already published four studiesMarie-Éléonore Marmion is project-team, which works on in this field.currently preparing a PhD in the modeling and parallel
  36. 36. the fruit of a passion“The path of a researcher is marked out by opportunities that can buildyour career. Ever since I was in high school, I have been passionateabout computer graphics. My plan was to go into the world of videogames or cinema. I didn’t even know that research in the field existed.Then I met various people, one thing led to another, and eventuallyI ended up doing research work that was awarded the Eurographicsprize, an international distinction that singles out contributions incomputer graphics. Receiving a prize is of course a great pleasure,but it is really a reward for teamwork and making the right choices.A researcher explores fields without ever knowing whether they willinterest anyone but himself; a prize brings the certainty that you werenot on the wrong track, that your work serves a purpose. But I didn’taim for the prize, it was simply a result of my passion!”Sylvain lefebvre joining Inria. From the start, the Eurographics prizeAfter a PhD thesis in computer he endeavoured to optimise in 2010, has had a considerablegraphics, Sylvain Lefebvre spent the computing of textures used impact on the academic as wella year in Seattle in the Microsoft in computer graphics. as industrial world.laboratories in 2005, before His work, which was awarded
  37. 37. Striving to improve living conditions“In September 2010, I attended a workshop on the use of new tech-nologies to help people with reduced autonomy. This event changedthe way I looked at public research missions. It prompted me notto make a complete U-turn, but to project my work towards a newpurpose: providing these people with technological tools that wouldimprove their quality of life. Applying my research to this social issue,and working with multidisciplinary teams that include researchersin computer science, psychology and cognitive sciences… the ideareally appealed to me. We therefore began to set up a number ofpartnerships, in particular with the association Trisomie 21, theUniversity of Bordeaux II, and the University of Quebec in Trois-Rivières (Canada).”Émilie balland diploma in mathematics and she is a research scientist inÉmilie Balland graduated from the computing. Her aim at the time the Phoenix project-team,University of Nancy in computer was to become a speech therapist. which focuses on programming-science, a subject she discovered Her discovery of computer science language technology forwhile studying for her university turned into a true passion. Today, communication services.
  38. 38. a memorable moment in the life of a researcher“All researchers hope one day to obtain their Habilitation to advisedoctoral theses, the highest educational qualification in France. Toprepare for it, researchers must define their own scientific fieldof expertise and put it in perspective. This preparation includes apsychological dimension, self-inspection both on a personal anda professional level. Some researchers feel ready after five years,others after twenty; personally, it took me nine years to take the step,a time during which I assisted many PhD students. Through mycontact with them, I became certain that my future would lie in mypreferred field, which is 3D interaction with virtual worlds. I receivedthe qualification in 2010, which enables me today to supervise thework of other future researchers, which is very motivating in itself.”anatole lÉcuyer to further studies. In 2001, aeronautics. He then joinedAfter graduating from the he defended a thesis on the Bunraku project-teamÉcole Centrale in Lille, Anatole tactile interaction with virtual to focus on virtual reality.Lécuyer first started a career worlds in industrialin engineering before returning maintenance operations for
  39. 39. Sharing knowledge: a mission and a pleasure“Presenting our research work to the general public is challengingand very satisfying at the same time. I recently tried my hand at it atthe Palais de la Découverte in Paris, as part of the “One researcher,one manipulation” campaign, on the topic of speaker recognition,in other words “knowing who’s talking” in a recording, based oneach person’s voice signature. Choosing the right words and pace toexplain a complex subject is a real challenge that requires carefulthought, since the methods implemented for voice authenticationrely on advanced mathematical knowledge. As for the satisfying part:that comes from the pleasure of accomplishing one of the essentialtasks of a researcher, which is to give the keys to understanding andto let each person form their own opinion. In addition, the questionsthat the audience ask also help us to reflect on what we do.”nancy bertin transcription. Besides her love (modeling and testing forA graduate from the ParisTech for mathematics, she also plays the processing of informationTelecom engineering school, the piano, the violin, and sings. and sound signals).Nancy Bertin did her PhD thesis She joined Inria in 2010 as parton automated music of the Metiss project-team
  40. 40. Working on an international level is very enriching“Before I joined Inria, I spent a year doing research at the Universityof Washington. That might seem commonplace nowadays, but in 1992it was quite an unusual thing to do. Ever since, I’ve always believedthat research is something to be done on an international scale, as thecombination of different training and cultural backgrounds acts as asource of inspiration. From the moment I became a researcher, I’ve beeninvolved in European projects, like Connect, which brings together tenEuropean partners (universities, organisations and companies) to focuson issues regarding network communication. Though the coordinationof international cooperation can be very time-consuming, managingteams in this kind of context is very motivating. If you have a coreof people who are used to working together, it creates the necessaryimpetus. In the end, the effort is always thoroughly rewarded by thequality of what is exchanged.”valÉrie iSSarny the Arles project-team to the start-up of Ambientik,Valérie Issarny graduated from (software architectures and a company that specialises inthe University of Rennes with distributed systems) which cooperative mobile applicationa PhD in computer science. She cooperates on international services, of which Valériejoined Inria where she manages projects. In 2011, her work led Issarny is a cofounder.
  41. 41. research driven by operational challenges“Today, we have models that can forecast air quality, as well as a varietyof new observation tools (satellites, microsensors). We have to meetthe challenge of learning how to use all these information sourcesin the best possible way. We are fortunate in that we cover a widespectrum, from the development of advanced mathematical methodsright through to their application in the software that we design. Forexample, together with the company Numtech and the associationAirparif, we are building a prototype that can estimate almost in realtime the exposure to pollution along an itinerary that a user in Parisdefines by himself on his mobile phone. This type of project allows usto match up our methods to real applications, identify new researchissues, and sometimes wind up with unexpected solutions.”vivien mallet nationale des ponts et chaussées. Among other things, he is inVivien Mallet graduated as an He joined Inria in 2007 as charge of the Polyphemus projectengineer from the Ecole Centrale a research scientist in the Clime to create a multimodel platformin Lyon, and obtained a PhD in project-team (modeling in for atmospheric pollution and riskapplied mathematics at the École environmental sciences). assessment.
  42. 42. 40/ InrIa annual report 2010
  43. 43. over the next pages, you will discover the life cycle of aproject-team (with an average duration of eight years). asan illustration, we look back at the first four years of theyoung Magrit team, and the last four years of the alchemyteam that is about to disperse.The Magrit project-team’s research focuseson augmented reality. led by Marie-odile Berger, Magrit is ajoint project-team between the French national Centre forScientific research (CnrS) and the universities of nancy.Set up in 2006, its activity was extended for another fouryears in 2010.The Alchemy project-team, a joint collaborationbetween the CnrS and the university of paris-Sud,studies architectures, languages and compilers for high-performance embedded or general processors. after eightyears of activity, notable results and the unequivocalsuccess of its applications, the alchemy team will nowpursue its work in new projects. InrIa annual report 2010 /41
  44. 44. topICS: perCeptIon, CoGnItIon, InteraCtIonThE MAGRIT PROjEcT-TEAM IN 2006n 2 InrIa reSearCHerS n 3 leCturerS n 5 pHD StuDentSn 1 poSt-DoCtoral reSearCHer n 2 enGIneerS OcTOBER 2006 a new perManent researcher for MaGrIt Development of the use of probabilistic methods in computer vision with the arrival in the team of Frédéric Sur, lecturer at the NOVEMBER 2006 École des Mines in Nancy. Best-paper award at IsMar Prize for the best paper at ISMAR (International Symposium on Mixed and Augmented Reality), on the influence of camera calibration errors on the quality of the augmented scene. 2006 MAGRIT PROjEcT-TEAMStart of the Magrit projectauGMenteD realItY IS a neW FIelD tHat HaS Been DeVelopeDConSIDeraBlY BY MaGrIt In tHe FIrSt Four YearS oF ItS eXIStenCe. since her thesis in 1989-1991, underBaCKGrounD anD HIGHlIGHtS. the supervision of Roger Mohr. Magrit’s goal is to develop researchThe history of Magrit starts in project-teams was recommended. In in augmented reality (AR),2006 with the end of the Isa the Vision group, which counted a discipline that aims to increaseproject-team (image, synthesis, five permanent researchers (two a person’s perception by addinganalysis) that was led by Inria researchers and three information to their field of visionJean-Claude Paul at the Inria university lecturers), five PhD to improve their understanding ofNancy – Grand-Est centre. students, one post-doctoral their environment. Marie-Odile BergerIsa at the time comprised more researcher and one engineer, recalls one of the first highlights:than 30 team members working Marie-Odile Berger, an Inria “To boost our work, we supportedin three distinct fields: computer researcher and university-qualified Frédéric Sur’s application for a lecturervision, computer graphics, and mathematics professor, decided to post. He is an expert in probabilisticgeometric visibility. During take the step of becoming a team methods, which is a very importantthe team’s evaluation in 2005, manager by forming Magrit. She part of automating the buildingthe creation of three new had been working on this topic of complex environment models.”42/ InrIa annual report 2010
  45. 45. topICS: arCHIteCture, proGraMMInG, CoMpIlatIon, lanGuaGeSThE ALchEMY PROjEcT-TEAM IN 2006:n 3 InrIa reSearCHerS n 1 leCturer 10 pHD StuDentS anD poSt-DoCtoral reSearCHerS OcTOBER 2006n aLcheMY InnoVates three joInt projects wIth Alchemy studies a new type the european unIon of computer architecture Milepost, Sarc and Acotes, described as neuro-inspired. three research projects that Alchemy This original research angle aims presented to European bodies, to design alternative architectures received considerable funding able to meet technological challenges in 2006 for a period of three years. encountered in the designing of These projects focus respectively processors, such as energy saving and on the design of efficient compilation possible defects in components. methods for complex architectures, To explore this route, Hugues Berry, the creation of new heterogeneous a biologist by training who specialises multiprocessor architectures, in the modeling of living things, and the development of feed was recruited by the team as programming methods for video. an Inria researcher in 2006. 2006 ALchEMY PROjEcT-TEAMalchemy: already a well structured team in 2006tHe appearanCe oF neW applICatIonS anD neW eleCtronIC CoMponentS CoMBIneD WItH tHe eMerGenCe oFConSIDeraBle reStrICtIonS (enerGY SaVInG, roBuStneSS) ConStantlY DrIVeS tHe DeVelopMent oF arCHIteCture,proGraMMInG, CoMpIlatIon anD lanGuaGe reSearCH. a SItuatIon eXperIenCeD BY tHe alCHeMY teaM.Alchemy’s history traces back to compilation and architecture,” to found Alchemy: bringing2003. The project-team was born explains Olivier Temam, then team architecture and compilationfrom the union between the A3 leader. From the start, Alchemy, researchers together in the fields ofteam that at the time counted two which had only just been set up in embedded systems and highpermanent researchers, Christine Saclay, became involved in a performance. But this time, the ideaEisenbeis and Albert Cohen, who large-scale European project that was to bring them together onworked on compilation, and the turned out to be pivotal, namely a European scale.”Architecture team run by Olivier HIPEAC (European Network ofTemam, professor at LRI. Why this Excellence on High Performancemerger? “We believed that most of the and Embedded Architecture andprogram performance problems in Compilation). “We took an activemodern architectures stemmed from a part in setting up this network withlack of communication between the same philosophy that prompted us InrIa annual report 2010 /43
  46. 46. FEBRUARY 2008 Second prize for a thesis augmented fluoroscopy, in 2008 awarded by the Lorraine an application of augmented region, on the concept of reality in the field of interventional neuroradiology. ARc 2007-2008 The research was conducted Took part in a collaborative jointly with the industry partner research action (ARC) GE Healthcare. It consists in conjunction with the teaching of superimposing pre-surgical 3D hospital in Nancy and the images onto per-surgical images Alcove project-team to enable neuroradiologists (Inria Lille–Nord Europe) studying to better target the desired area the simulation of coil rollout (aneurism) in order to insert a in the treatment of cerebral stent, a cuff, etc. aneurism. PhD thesis award for the Lorraine region. 2007 MAGRIT PROjEcT-TEAMFundamental research oninteractive modelingMagrit aims to develop solutions calculated at every instant,” possible to obtain structured modelsfor calculating insertion and explains Marie-Odile Berger. of the scene, and on the other hand,visual reconstruction, two of “The other cornerstone is the 3D control the quality of the reconstructedthe main challenges to be tackled reconstruction of the environment models in real time by comparingto enable potential applications a person sees.” Modeling is essential them with real vision. The idea isof augmented reality to become here, for example to take into account to design interactive modes that areoperational in the long run, the interaction of light between simple for the user, and which offerin large spaces. At the moment, (added) virtual and real (on the scene) optimal reliability.” An academicmost of the applications only work objects.” research field that lies betweenin limited spaces and over short “To address these fundamental issues, the “vision” community andperiods of time. “To integrate we studied fully automated methods. the “computer graphics”information at the right spot in Since 2008, we have also been community which Magrit hasthe field of vision, regardless of studying interactive methods that let developed to a great extent thesethe movement the user makes, the user participate in the application. past few years.the observer’s point of view must be These on the one hand make it44/ InrIa annual report 2010
  47. 47. 2007 aLcheMY consoLIdates aLcheMY: IMportant IndustrIaL Its work on IteratIVe partnershIps reGardInG the coMpILatIon Methods “proGraMMInG” aspect In 2007, the Alchemy team’s Direct contracts as well as Cifre recruitment of Grigori Fursin, contracts have been signed with: who holds a PhD from Edinburgh – Hewlett Packard France on the University, allowed them to notion of programming optimisation develop particular compilation (2004-2007); methods, known as iterative – Philips (now NXP) on synchronous language compilation methods. This new programming approaches (2000-2009); approach makes it possible to – ST Microelectronics on programming adapt compilers to complex aspects for complex microprocessor architectures. The work led to architectures (2006-2010). the design of a smart compiler All of these relationships continue (Milepost GCC) in partnership with today in other research fields related to IBM Research. programming and architecture. 2007 ALchEMY PROjEcT-TEAMCompilation/architecture interaction,at the heart of alchemy’s researchThe idea of creating interaction to recruit Grigori Fursin. Hisbetween compilation and mission was to develop iterativearchitecture has come a long way… compilation techniques. From 2006including with regard to European to 2009, the team’s researchauthorities. The team’s involvement activities focused on compilation/in the HIPEAC network has had architecture interaction to meet thenumerous consequences. Funding objectives of the three European Over the next pages, follow the Alchemy team with Christine Eisenbeis (p.47) andfor the three research projects projects. “This meant that we had less Olivier Temam (p. 43).(Milepost, Sarc and Acotes) was for leeway to concentrate on the other areainstance obtained from the that we wanted to develop in Alchemy,European Commission: a total of which is the growing complexity in800,000 euro per year for three years technological development (Moore’s(2006-2009). This made it possible to law) and the impact thereof,” explains which stands for Architectures,receive a number of PhD students Oliver Temam. “This refers to the Languages and Compilers to Harnessand post-doctoral researchers, and second part of the acronym ‘Alchemy’, the End of Moore Years.” InrIa annual report 2010 /45
  48. 48. 2009 Best paper award at the IccV euroGraphIcs 2009, MunIch a new perManent 2009 workshop In kYoto During this European conference researcher for MaGrIt “On Video Oriented Objects and on computer graphics, Magrit Magrit recruits Pierre Fréderic Event Classification”: the results presented an interactive in situ Villard, a university lecturer. This related to the modeling and model-building method using specialist in physical modeling recognition of operational phases a camera. formerly worked as a post- (based on a series of videos taken doctoral researcher at Imperial in an operating theatre). They College, London. His arrival on were obtained in the scope of a the team has enabled them to thesis co-supervised by Magrit develop an aspect that they and the Technical University of had previously left relatively Munich (TUM). unexplored, regarding the construction of physical models, essential for obtaining realistic simulations. 2009 MAGRIT PROjEcT-TEAMMedical applications foraugmented realityApplications in particular include research activities on medical built on this work over the yearscomputer-assisted medical imaging, in particular and today the team’s researchinterventions and the designing of interventional radiology. In 2007, forms part of the large-scale action,surgical simulators. Since it came Erwan Kerrien got in touch with Sofa InterMeds. Certain medicalinto being, Magrit has focused its Stéphane Cotin of the Alcove applications are being studied in project at the Inria Lille – Nord conjunction with industrial Europe centre. Their collaboration partners: several theses have for concerns the simulation of coil instance been completed in rollout in blood vessels (see page 44), collaboration with GE Healthcare and the modeling of vessels from in the scope of Cifre contracts, 3D images (angiograms). Their aim particularly on augmented is to obtain faithful, efficient fluoroscopy. representations of the vascular network in order to simulate operations in real time. Magrit has46/ InrIa annual report 2010
  49. 49. 2009 desIGn of MILepost Gcc, a sMart open-source coMpILer coLLaBoratIon wIth chIna This compiler is today widely available Cooperation with the academic institute to the public. It was co-designed for sciences of China (ICT) in Beijing, in partnership by Alchemy, the University on iterative compilation and the combined of Edinburgh, and IBM Research in language/architecture approach. the scope of the Milepost project. ICT is the research centre in charge It includes results obtained on iterative of designing the Chinese Loongson compilation methods developed by processor. This cooperation will G. Fursin and O. Temam. Milepost GCC be consolidated in 2011 with the was designed to optimise compilation creation of an associate team, Youhua in an automated way. Other results that (“optimisation” in Chinese). relate to the polyhedral model (a new way of representing programs in a compiler) and that were studied by Albert Cohen were introduced in the public GCC compiler.2009 ALchEMY PROjEcT-TEAMarchitecture and programming: of new applications, increasing constraints in terms of energydealing with future constraints consumption and tolerance to defects, and lastly, the development of new According to Moore’s law, the size electronic components (memristors), of transistors in chips is halved gradually led us to consider neuro- every two years. This law has been inspired architectures,” explains followed for more than thirty years. Olivier Teman. “To explore this “It was still valid when we created direction, we recruited Hughes Berry, Alchemy, but we realised that it was a biologist who specialises in the being exposed to enormous pressure modeling of neurons, in 2010.” (excessive energy consumption and This line of research had previously defects due to process technology), remained somewhat in the and we wanted to explore alternative background, since most of the routes.” Hence the need to develop team’s researchers had been working new architectures and programming on European projects – a state methods that are more energy- of affairs that changed in 2009, efficient and tolerant to defects when the European contracts came in components. “The appearance to an end. InrIa annual report 2010 /47