The proposed research programme aims at identifying noise-generating mechanisms in subsonic turbulent jets, and at the development of closed-loop control laws for the reduction of jet noise through flow actuation. An interdisciplinary approach combines experiment, numerical simulation and theoretical modelling in a coordinated effort, between three partner institutions with complementary expertise. While optimal control laws can, in principle and at enormous computational cost, be devised on the empirical basis of numerical simulations, taking into account the entire turbulent spectrum, the present proposal focuses on the dominant noise component associated with large-scale coherent flow structures, that drive the low-angle sound field. Fundamental progress in the understanding of the dynamics of these coherent structures, as well as their sound generation, will provide guidance for novel strategies to actively control and reduce jet noise. The programme addresses the following questions: Which mechanisms govern the formation of orderly structures in jet turbulence? Can these structures be accurately described as instability wavepackets forming on top of a steady mean flow, as has often been conjectured? To what extent do nonlinear phenomena determine the wavepacket structure and the resulting acoustic field? And how can knowledge of these mechanisms be leveraged for jet noise reduction? Control strategies will be devised, and these will be tested in a real experiment during the final stage of the project. The proposal builds on ongoing research activities at the three partner institutions, which so far have been developed independently without formal collaboration. The synergy potential of these complementary activities is considerable, and the proposal precisely aims to provide a framework for a coordinated interaction with a common set of objectives. Operational tools and preliminary results exist for all the main stages of the proposed programme. These include ongoing experiments on jet dynamics and their acoustic signature at PPRIME; a validated LES code; numerical tools for jet instability analysis at LadHyX, that are currently used on model configurations and await application on real-life jet data; model-free control concepts, developed at LadHyX, ONERA and LIMSI, that have been successfully deployed to reduce sound emission from flow over cavities; and reduced-order modeling for flow control (ANR Chair of Excellence at Pprime). International collaborations on jet noise research, with Tim Colonius at the California Institute of Technology and with André Cavalieri at Instituto Tecnológico de Aeronáutica (Sao José dos Campos, Brazil), are already in place and will be further intensified during the course of the proposed programme. The proposal seeks funding for (i) one PhD student (3 years) and four postdoc years; (ii) experimental equipment for particle image velocimetry in high-speed jets; (iii) travel expenses for conference participation and for the collaboration between partners, including the external collaborators at Caltech and at ITA.

Situated on the Northern Fringe of the Massif Central (France), the Linguistic Crescent encompasses parts of the following French departments: Eastern Charente, the Southern edges of Vienne, Indre and Cher, the Northern margins of Haute-Vienne, Creuse and Puy de Dôme, and South Allier). On a map, this strip of land has the shape of a half-moon, hence its name (first used by Ronjat in 1913). The Crescent is an area where the local gallo-romance varieties simultaneously display typical Oïl (French, Poitevin-Saintongeais, Berrichon) and Occitan (Limousin and Auvergnat) features. Today, these local varieties remain largely underdescribed, mainly due to their dual nature, which has made the specialists of the two main gallo-romance entities which adjoin the Crescent (i.e. Oïl in the North and Occitan in the South) equally reluctant to include them in their respective areas of expertise. The aim of this project is to provide a multidisciplinary approach to these now endangered varieties (whose majority of speakers are above 70 years old) with the following main lines of research : (1) the building of a multi-dialectal corpus comprising with (i) oral texts and (ii) video documents collected all across the Crescent area and stored in accordance with the usual practices in today's documentary linguistics (data transcribed in ELAN, metadata in ARBIL, systematic use of language-processing methods); (2) a comparative approach to the Crescent varieties, covering several tens of investigation points, including both phonological, lexical and morphological data and using a reference database in combination with mapping applications; (3) description (fieldwork) and typological analysis (especially in the domains of acoustics, prosody, morphology and formal semantics) resorting to promising, up-to-date methodological and theoretical frameworks (e.g. processing of audio data through Praat or similar software, modeling of verb paradigms by means of NetLog); (4) characterization with regard to neighboring (oc/oïl) varieties; (5) psycholinguistic studies of French/local language bilingual speakers by testing the sensitivity of these speakers to different parameters such as stress, vowel length and quality ; (6) sociolinguistic study of the representations associated with the Crescent varieties. These actions are designed to document and enhance before it is too late as much as possible of this rich, largely unexplored linguistic heritage by making available open-access data (available through HumaNum), high-level scientific publications as well as helping local communities to take advantage of their own cultural resources (the project will be developed in close partnership with cultural associations established in the Crescent area and will actively support the local dissemination of the scientific results through meetings and other appropriated events). The project team includes many of today's specialists of Crescent varieties; in addition to these, it also includes other researchers enjoying a high degree of recognition in their respective fields of expertise and who have decided to devote some of their research time to the Crescent, due to the particular interest of this topic for their own investigations.

This project is a trans-disciplinary effort toward inference and prediction of complex physical systems. Such systems are often ineffectively described by first principles models and should be modeled via a data-driven approach. However, difficulties arise from the high dimensional and multi-scale nature of these systems. Further, only limited and poorly informative observations are typically available. Prototypical of these situations is subsurface ocean inference or the prevention of seizures in neurosciences. For many applications however, some degree of expertise is available. The goal of this project is to leverage both the theoretical and the data science pillars to infer computable models informed from the existing prior knowledge (Physics first principles and theories) and providing new hints into the principles satisfied by the proposed abstractions, amenable to interpretation and refutation. More precisely, building upon the pluridisciplinary expertise of the team in the domains of Fluid Mechanics and Deep Neural Networks (DNN), the goal of the proposal is: - i) to make the model space (neural architecture and computational flow) compliant with the known Physics of the system under consideration, - ii) to exploit the data and inference tools to train efficient models built on first principles, thereby enhancing their robustness and reducing their data-hunger, - iii) to form and inspect the abstractions built by the DNN systems, to check whether they satisfy the expected properties and understand the properties they satisfy. Methods and tools will be first developed with low-dimensional dynamical systems but will then be illustrated and demonstrated on a full-scale turbulent fluid flow numerical simulation.

The goal of the SALSA project (Speech And Language technologies for Security Applications) is to develop a set of speech and language processing tools specifically designed to assist analysts in processing and exploiting audio data for security purposes, such as judicial, law enforcement and intelligence applications. The SALSA consortium is composed of 4 technology and research partners with complementary expertise and excellent track records in their respective fields, and 3 user partners, also members of the advisory board. The coordinator, Vocapia Research (www.vocapia.com) is a software editor specialized in speech processing. The Centre National de la Recherche Scientifique - Laboratoire d'Informatique pour la Mécanique et les Sciences de l'Ingénieur (CNRS-LIMSI) brings extensive expertise in spoken language processing. Laboratoire de Phonétique et de Phonologie (LPP), a CNRS-Paris 3 University research unit, brings acoustic-phonetic and articulatory expertise of speech production from a linguistic perspective. The fourth partner, Intelligences, specializes in providing expertise in using language technologies for judicial investigations. The three user partners are the French Defence Procurement Agency (DGA), the General Directorate of the National Police (DGPN) and the Home Affairs Technologies and Information Systems department (ST(SI)2) which is also federating the needs of several other agencies (DCPJ, IRCGN, OCRIEST, PP and PTS). These users will ensure that SALSA covers the spectra of needs that security agencies face when coping with large amounts of speech data. Filtering information in audio data is critical for national security to ensure the protection of citizens and is of strategic need for many governmental agencies. The SALSA project will improve over the current state-of-the-art in language technologies in order to develop aids for analysts currently unable to process the exponentially growing volumes of data. The main objectives of the project are to enhance the efficiency of the analysts while reducing their workload, and to provide support for novel data mining in audio. To do so, the following technological innovations will be explored: New learning methods for spontaneous speech; Linguistic investigations of accented speech and speech with code-switching; New decoding methods for transcription and keyword spotting; and Novel user interfaces for intelligence analysis of audio data. The three users will be involved in the development loop and evaluation ensuring a close match to the security analysts needs.

From the original idea to the actual distribution on TV, VOD and DVD, the production process and subsequent “life” of a TV program are divided into numerous stages, involving numerous actors (e.g. distribution, production, direction, screenwriting, casting, post-production or dubbing) and thus leading to the generation of a huge amount of heterogeneous metadata. However, only a few metadata eventually survive the tortuous production and distribution processes, making their integration difficult into novel TV-centric products. Even for TV productions where one actor manages the whole production pipeline (e.g. Canal+ in France or BBC in the UK), most of the metadata do get lost at one point or another. The MetaDaTV network proposal aims at initiating a European research community around metadata associated with TV productions (such as dramas, documentaries or TV films) and at gathering interested partners from all over Europe toward a joint European project submission (Horizon 2020).