Description du projet
C’est maintenant que se joue l’avenir des technologies d’interaction vocale
Les technologies relatives aux commandes vocales ont fait de grands progrès ces dernières années. À vrai dire, la voix est en train de remplacer rapidement le toucher et le texte, en tant que principal moyen d’interaction avec les appareils modernes. C’est dans ce contexte que le projet COMPRISE, financé par l’UE, développera la technologie d’interaction vocale de prochaine génération. Son objectif consiste à réduire les coûts liés à la collecte et à l’étiquetage des données vocales pour toute une variété de langues, afin de rendre cette technologie plus inclusive. Actuellement, les technologies d’interaction vocale privilégient l’anglais et d’autres langues disposant d’une vaste base d’utilisateurs. La viabilité de ce nouvel écosystème, qui proposera ses services à des locuteurs s’exprimant avec un accent marqué ou dans une langue sous-équipée en termes de ressources, fera l’objet de démonstrations dans trois secteurs à fort impact commercial: les applications intelligentes pour les consommateurs, le commerce électronique et la santé en ligne.
Objectif
Besides visual and tactile, the Next Generation Internet will rely more and more on voice interaction. This technology requires huge amounts of speech and language data in every language to reach state-of-the-art performance. The standard today is to store the voices of end users in the cloud and label them manually. This approach raises critical privacy concerns, it limits the number of deployed languages, and it has led to market and data concentration in the hands of big non-European companies such as Google, Facebook, etc.
COMPRISE defines a fully private-by-design methodology and tools that will reduce the cost and increase the inclusiveness of voice interaction technology through research advances on privacy-driven data transformations, personalised learning, automatic labelling, and integrated translation. This leads to a holistic easy-to-use software development kit interoperating with a cloud-based resource platform. The sustainability of this new ecosystem will be demonstrated for three sectors with high commercial impact: smart consumer apps, e-commerce, and e-health.
COMPRISE will address the mission-oriented challenges of privacy-by-design, inclusiveness, and cost-effectiveness in a sector-agnostic way; allow virtually unlimited collection of real-life non-private quality speech and language data; enable businesses in the Digital Single Market to quickly develop multilingual voice-enabled services in many languages; allow all citizens to transparently access contents and services available in other languages by voice interaction in their own language; result in cost savings for both technology providers and users.
COMPRISE will find application in many sectors beyond those demonstrated, e.g. e-government, e-justice, e-learning, tourism, culture, media, etc. It will have a huge societal impact in terms of unprecedented verifiable privacy guarantees, service to speakers of under-resourced languages or accented speakers, and overall user experience.
Champ scientifique
Not validated
Not validated
- medical and health scienceshealth scienceshealth care serviceseHealth
- natural sciencescomputer and information sciencesinternet
- natural sciencescomputer and information sciencessoftwaresoftware development
- social scienceseconomics and businessbusiness and managementcommercee-commerce
- social sciencespolitical sciencesgovernment systemse-governance
Mots‑clés
Programme(s)
Régime de financement
RIA - Research and Innovation actionCoordinateur
78153 Le Chesnay Cedex
France