The French-based mega technology project shows some small signs of progress, but critics say its top-heavy, public structure stands in the way of any real innovation
PARIS - What exactly is Quaero searching for? It's a 200-million-euro question, in light of the huge investment in this research program, which was selected by the Agency for Industrial Innovation and launched by then-French President Jacques Chirac. Since its birth in 2006, Quaero has been called the ‘European anti-Google" and has been the source of repeated criticisms: it is too big, too expensive, illegitimately controlled by French industrial group Thomson -- and ultimately destined for failure.
Since its founding, Quaero's German industrial partners have left the venture, and Thomson has become Technicolor – and it took two years to even give a concrete existence to the project. It effectively launched in 2008, and is expected to be completed in 2013. After nearly three years, the 26 program partners maintain a low profile, no longer mention Google, and let their results speak for themselves: twenty patent applications, nearly 300 scientific publications, and a multitude of applications and software programs.
"Quaero is a research accelerator that helps to stimulate the transfer of technology to industry," says Pieter van der Linden, research director at Technicolor and coordinator of the program. Nevertheless, the precise contours of the project, which aims to be "the first European center for research and development of the automatic extraction of information, analysis, classification, and the use of multimedia and multilingual content" are difficult to identify. Even for specialists.
"Quaero brings together a team of some 30 partners, and it has been difficult to navigate, even for us," says Olivier Galibert, an engineer at the Laboratoire National d'Essais (LNE), one of the agencies responsible for assessing Quaero.
Specifically, Quaero has created a "core technology cluster" (CTC), which is led by the French government-funded National Center for Scientific Research (CNRS) and the German RWTH Aachen research university. Nearly half of the 300 Quaero researchers and engineers work for this CTC, which is "an R&D tool shared by the entire program," explains van der Linden. Quaero is organized around five application projects that are each coordinated by industrial leaders: Jouve, France Telecom, Exalead, Yacast and Technicolor.
Exalead (owner of the search engine of the same name) is leading a project focused on the exploitation of multimedia documents. One of its primary achievements has been Voxalead, a multimedia search engine that implements Vocapia technology, which was developed by the LIMSI-CNRS Laboratory outside of Paris. Voxalead, an easy-to-use research software, uses keywords to find passages or radio news programs that correspond to any given topic.
Proponents of this tool are considering multiple applications, from the automated analysis of recorded conversations in call centers to the expansion of e-learning. "The application has already been sold to KSU University in Saudi Arabia, where all classes are videotaped and available online," says Jean-Marc Lazard, who is in charge of strategic projects at Exalead.
Measuring Speaking Time
Yacast, which specializes in audiovisual media analysis (radio music monitoring, measurement of advertising investment), has been supervising projects around the clock. One system, media monitoring social impact (MMSI) is used by Radio France to measure the speaking time of politicians. "With Quaero, we worked a lot on speaker identification, which allows us to automate this process," explains David Scravaglieri, Director of R&D at Yacast. The interface allows the company to monitor speech counts by candidate and by political party. MMSI is now also proposing to link with the blogosphere, in order to judge the effects of political discourse on blog content.
Another project concerns the personalization and interactivity of television content. One of the software platforms developed through Quaero allows users to browse an online video store that proposes selections based on their user profiles and past purchases via a gyroscopic remote control that controls a virtual environment in the home. "Imagine that you can order your evening movie while still watching the news on the big screen," explains Thierry Filoche, Project Manager at Technicolor.
"A very state-based approach"
Despite these projects, Quaero has become a punching bag for others in the high-tech industry. "Quaero is a scandal," says Jean-Pierre Gérault, President of the Richelieu Committee and CEO of I2S, which runs Polinum, a consortium for the digitization of heritage collections. "It's a good idea, but has been poorly implemented. To put a large group in charge of this kind of project is a mistake. Large companies are not inherently collaborative, and it winds up as a waste of money."
Stéphane Distinguin, the founder FaberNovel, a start-up specializing in interactive content, is also critical of Quaero: "This project started with a very state-based approach. It was very arrogant and will be costly to the (high-tech) community, especially to small projects. I imagine we will see very little from it in terms of disruptive innovation."
The specialists in charge of assessing Quaero have thus far remained calm. "It is true that the costs are high and the duration of the project is long, but this will allow it to go further" than other public research projects, maintains Olivier Galibert. In response to another recurrent criticism, the relative secrecy enshrouding the project, the Quaero team has promised to offer a detailed progress report of their latest projects this fall.
Read the original article in French