Lieu : Orléans/Grenoble, France
Contacts : Emmanuel Schang (emmanuel.schang@univ-orleans.fr), Benjamin Lecouteux (benjamin.lecouteux@univ-grenoble-alpes.fr)


Nous cherchons un?e candidat?e pour une thèse en Sciences du Langage sur le thème du traitement automatique de la parole.
La thèse s?effectuera au sein du Laboratoire Ligérien de Linguistique (LLL, UMR 7270), avec une possibilité d'accueil au LIG-GETALP (Grenoble). Le financement se fera dans le cadre du projet ANR CREAM (Documentation des langues CREoles Assistée par la Machine https://sites.google.com/view/creamproject/home).

Termes-clés : langues créoles, traitement automatique de la parole, détection de mot clé, alignement bilingue, creole languages, speech processing, keyword spotting, bilingual alignment.


Objectifs
Le projet CREAM  vise à proposer aux  linguistes travaillant sur les langues créoles des outils novateurs dans la collecte et le traitement des données orales sur des langues disposant de peu de ressources.
Dans le contexte particulier de diglossie qui caractérise souvent l'espace créolophone, le passage par l'étape de la transcription de corpus est fréquemment ressenti comme une difficulté par les linguistes de terrain. Une conséquence est le manque de corpus disponibles.

L'objectif de ce projet est d'ouvrir la voie à des méthodes novatrices en matière de documentation linguistique et de création de ressources sur les langues créoles. En utilisant des technologies d'apprentissage automatique de pointe, nous cherchons à changer la façon dont la documentation linguistique est mise en ?uvre en termes de construction de ressources linguistiques et de traitement des corpus parlés.

L'accent sera mis sur deux tâches en particulier :
- Query-by-example : la recherche de segments similaires dans des corpus en langue créole,
- Alignement bilingue automatisé entre des segments de parole dans une langue créole et une langue proche (français, anglais, portugais, suivant les créoles).

Selon les avancées, les recherches pourront s'étendre à d'autres tâches du TAL :

- la reconnaissance automatique de la parole (étude du transfert d'apprentissage entre langues lexificatrices et langues créoles)

- traduction automatique ...

Bibiographie sélective
- G. Adda, et al.. (2016). Breaking the unwritten language barrier: the BULB project. In SLTU-2016 5th Workshop on Spoken Language Technologies for Under-resourced languages, Yogyakarta, Indonesia, May 2016.
- A. Baevski, et al.  (2020). wav2vec 2.0: A framework for self-supervised learning of speech representations. arXiv preprint arXiv:2006.11477.
- D. Blachon, et al. (2016). Parallel Speech Collection for Under-resourced Language Studies Using the Lig-Aikuma Mobile Device App. In Workshop on Spoken Language Technologies for Under-resourced Languages (SLTU), Yogyakarta, Indonesia, May 2016.
- P. Godard et al. (2018). Unsupervised Word Segmentation from Speech with Attention. In Interspeech 2018, Hyderabad, India, September 2018.
- Y.-A. Chung, et al. (2016). Audio word2vec: Unsupervised learning of audio segment representations using sequence-to-sequence autoen- coder, Interspeech 2016 pp. 765?769.
- H. Kamper, (2019). Truly unsupervised acoustic word embeddings using weak top-down constraints in encoder-decoder models, ICASSP 2019-2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), IEEE, pp. 6535?39.
- H. Kamper, Anastassiou, A. and Livescu, K. (2019). Semantic query-by-example speech search using visual grounding, ICASSP 2019-2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), IEEE, pp. 7120?24.
- S. Schneider, et al.  (2019). Wav2vec: Unsupervised Pretraining for Speech Recognition. Interspeech. Graz, Austria, 2019.
- V. Velupillai. (2015). Pidgins, creoles and mixed languages. John Benjamins Publishing Company.

Profil recherché
Les candidat.e.s auront un master en linguistique ou en informatique et montreront un intérêt certain pour le traitement automatique de la parole et les langues dites 'rares'. Une autonomie dans le codage en python est indispensable, ainsi que des bases en machine learning.

Candidature : les candidat.e.s enverront une lettre de motivation ainsi qu?un CV détaillé. Des documents complémentaires pourront être demandés si le ou la candidat.e est retenu.e pour une audition.

Encadrement
Emmanuel SCHANG (Docteur HDR en Sciences du Langage)
Benjamin LECOUTEUX (Docteur en Informatique)


Candidature à envoyer à Emmanuel Schang (emmanuel.schang@univ-orleans.fr), Benjamin Lecouteux (benjamin.lecouteux@univ-grenoble-alpes.fr).

Calendrier :

Date limite d?envoi des dossiers : 05 novembre 2021

Les dates des auditions seront communiquées aux candidat.e.s retenu.e.s sur dossier.

Top

Three fully-funded PhD positions on early prosodic development at Utrecht University

These positions are part of a new Dutch Research Council project ‘SoundStart’ on early prosodic development led by Professor Aoju Chen at Utrecht University. Children show knowledge of prosody (i.e., melody and rhythm) of their native language already at birth and thrive on it in early language and communication development. However, how children learn prosody so early is still unknown. Taking an interdisciplinary and crosslinguistic approach, SoundStart aims to discern the role of innate mechanisms, uncover learning mechanisms in the auditory modality underlying development spanning prenatal and after-birth periods, and shed light on the role of visual cues to prosody (i.e., speech-accompanying gestures) in after-birth periods. PhD projects 1 and 2 will be concerned with the first two goals in the learning of prosodic phrasing (i.e., grouping sequences of sounds into meaningful units in speech streams) and prosodic form-meaning mappings (i.e. associating prosodic patterns with communicative functions), PhD project 3 with the third goal in both areas of prosodic development. We are seeking to hire highly motivated, driven and talented MA graduates to take on the PhD projects, starting 1 February 2022. A later starting date is negotiable.

In this research programme you will work within an interdisciplinary team, closely collaborating with researchers from linguistics, neuroscience, psycholinguistics, psychology and neonatology at Utrecht University and the University Medical Centre Utrecht. You will receive assistance from student assistants in data collection (in the Netherlands and abroad) and/or data processing and have the opportunity to develop your academic teaching skills during your project.

Do you want to play a key role in this new exciting research programme on early prosodic development? Read more and apply! The application deadline is 8 November, 2021.

https://www.uu.nl/en/organisation/working-at-utrecht-university/jobs/three-fully-funded-phd-positions-on-early-prosodic-development-at-utrecht-university-10-fte

For more information about these positions, please contact Professor Aoju Chen at aoju.chen@uu.nl.

----

Prof. dr. Aoju Chen

Chair of Language Development in Relation to Socialisation and Identity

Head of the English Language and Culture Division 

Leader of VICI research group SoundStart

Department of Languages, Literature and Communication & Utrecht Institute of Linguistics - OTS

Utrecht University 

The Netherlands

Top

L'équipe Multispeech d'Inria Nancy recrute un ingénieur permanent dont la mission principale sera de piloter et contribuer au développement d'un assistant virtuel multimodal évolutif, à partir de briques logicielles open source et/ou issues de l'équipe, à des fins de recherche et de transfert industriel.

Fiche de poste: voir pages 9-10 du fichier https://www.inria.fr/sites/default/files/2021-10/Fiches_poste_ingenieurs_permanents.pdf  

Candidature en ligne: https://www.inria.fr/fr/concours-externes

Date butoir: 31/10/2021.

Top

L’équipe SAMoVA de l’IRIT à Toulouse propose plusieurs stages en 2022 :

- Analyse des signaux IoT (audio et accéléromètre) issus du collier Sallis Médical en vue de la modélisation de l’efficacité pharyngolaryngée
- Discriminative sequence training in end-to-end automatic speech recognition
- Le traitement automatique de la parole et l’IA au service du diagnostic différencié de maladies neurodégénératives
- Self- and semi-supervised adaptation of neural speaker diarization
- Self-supervised audio representation learning
- ...

Tous les détails (sujets, contacts) sont disponibles dans la section 'Jobs' de l’équipe, onglet 'Internships' :
https://www.irit.fr/SAMOVA/site/jobs/

Top

====== Offre de stage en reconnaissance automatique de la parole ========

Top

The SteelSeries France R&D team (former Nahimic R&D team) is glad to open 4 research internship positions for 2022.

Internship position at Telecom-Paris on Deep learning approaches for social computing

           

*Place of work* Telecom Paris, Palaiseau (Paris outskirt)

 

*Starting date* From February 2021(but can start later)

 

*Duration* 4-6 months

 

*Context*

The intern will take part in the REVITALISE project, funded by ANR.

The research activity of the internship will bring together the research topics of Prof. Chloé Clavel [Clavel] of the S2a [SSA] team at Telecom-Paris– social computing [SocComp] - and Dr. Mathieu Chollet [Chollet] from University of Glasgow – multimodal systems for social skills training, and Dr Beatrice Biancardi [Biancardi] – Social Behaviour Modelling from CESI Engineering School, Nanterre.

 

* Candidate profile*

As a minimum requirement, the successful candidate should have:

• A master degree in one or more of the following areas: human-agent interaction, deep learning, computational linguistics, affective computing, reinforcement learning, natural language processing, speech processing

• Excellent programming skills (preferably in Python)

• Excellent command of English

• The desire to do an academic thesis at Telecom-Paris after the internship

 

*How to apply*

The application should be formatted as **a single pdf file** and should include:

• A complete and detailed curriculum vitae

• A cover letter

• The contact of two referees

The pdf file should be sent to the two supervisors: Chloé Clavel, Beatrice Biancardi and Mathieu Chollet: chloe.clavel@telecom-paris.fr, bbiancardi@cesi.fr, mathieu.chollet@glasgow.ac.uk

 

           

Multimodal attention models for assessing and providing feedback on users’ public speaking ability

 

*Keywords* human-machine interaction, attention models, recurrent neural networks, Social Computing, natural language processing, speech processing, non-verbal behavior processing, multimodality, soft skills, public speaking

 

*Supervision* Chloé Clavel, Mathieu Chollet, Beatrice Biancardi

 

*Description* Oral communication skills are essential in many situations and have been identified as core skills of the 21st century. Technological innovations have enabled social skills training applications which hold great training potential: speakers’ behaviors can be automatically measured, and machine learning models can be trained to predict public speaking performance from these measurements and subsequently generate personalized feedback to the trainees.

The REVITALISE project proposes to study explainable machine learning models for the automatic assessment of public speaking and for automatic feedback production to public speaking trainees. In particular, the recruited intern will address the following points:

-   identify relevant datasets for training public speaking and prepare them for model training

-   propose and implement multimodal machine learning models for public speaking assessment and compare them to existing approaches in terms of predictive performance.

-   integrate the public assessment models to produce feedback a public speaking training interface, and evaluate the usefulness and acceptability of the produced feedback in a user study

The results of the project will help to advance the state of the art in social signal processing, and will further our understanding of the performance/explainability trade-off of these models.

 

The compared models will include traditional machine learning models proposed in previous work [Wortwein] and sequential neural approaches (recurrent networks) that integrate attention models as a continuation of the work done in [Hemamou], [BenYoussef]. The feedback production interface will extend a system developed in previous work [Chollet21].

 

Selected references of the team:

[Hemamou] L. Hemamou, G. Felhi, V. Vandenbussche, J.-C. Martin, C. Clavel, HireNet: a Hierarchical Attention Model for the Automatic Analysis of Asynchronous Video Job Interviews.  in AAAI 2019, to appear

[Ben-Youssef]  Atef Ben-Youssef, Chloé Clavel, Slim Essid, Miriam Bilac, Marine Chamoux, and Angelica Lim.  Ue-hri: a new dataset for the study of user engagement in spontaneous human-robot interactions.  In  Proceedings of the 19th ACM International Conference on Multimodal Interaction, pages 464–472. ACM, 2017.

[Wortwein] Torsten Wörtwein, Mathieu Chollet, Boris Schauerte, Louis-Philippe Morency, Rainer Stiefelhagen, and Stefan Scherer. 2015. Multimodal Public Speaking Performance Assessment. In Proceedings of the 2015 ACM on International Conference on Multimodal Interaction (ICMI '15). Association for Computing Machinery, New York, NY, USA, 43–50.

[Chollet21] Chollet, M., Marsella, S., & Scherer, S. (2021). Training public speaking with virtual social interactions: effectiveness of real-time feedback and delayed feedback. Journal on Multimodal User Interfaces, 1-13.

 

Other references:

[TPT] https://www.telecom-paristech.fr/eng/ 

[IMTA] https://www.imt-atlantique.fr/fr

[SocComp.] https://www.tsi.telecom-paristech.fr/recherche/themes-de-recherche/analyse-automatique-des-donnees-sociales-social-computing/

[SSA] http://www.tsi.telecom-paristech.fr/ssa/#

[PACCE] https://www.ls2n.fr/equipe/pacce/

[Clavel] https://clavel.wp.imt.fr/publications/

[Chollet] https://matchollet.github.io/

[Biancardi] https://sites.google.com/view/beatricebiancardi

-Rasipuram, Sowmya, and Dinesh Babu Jayagopi. 'Automatic multimodal assessment of soft skills in social interactions: a review.' Multimedia Tools and Applications (2020): 1-24.

-Sharma, Rahul, Tanaya Guha, and Gaurav Sharma. 'Multichannel attention network for analyzing visual behavior in public speaking.' 2018 IEEE Winter Conference on Applications of Computer Vision (WACV). IEEE, 2018.

-Acharyya, R., Das, S., Chattoraj, A., & Tanveer, M. I. (2020, April). FairyTED: A Fair Rating Predictor for TED Talk Data. In Proceedings of the AAAI Conference on Artificial Intelligence (Vol. 34, No. 01, pp. 338-345).

Top

Hi,

Top

Afin de renforcer son équipe, le laboratoire Commun ALAIA, destiné à l'Apprentissage des langues Assisté par Intelligence Artificielle, propose deux postes d'ingénieurs de recherche (12 mois).

ALAIA est centré sur l'expression et la compréhension orale d'une langue étrangère cible (L2). En collaboration avec ses deux partenaires, académique (IRIT) et industriel (Archean Technologie) ainsi que des experts en didactiques de langues, les missions consisteront à concevoir, développer et intégrer des services innovants basés sur l'analyse des productions des apprenants L2, la détection et la caractérisation d'erreurs allant du niveau phonétique au niveau linguistique. Elles seront affinées en fonction du profil des personnes recrutées.

Les compétences attendues portent sur le traitement automatique de la parole et du langage ainsi que les méthodes de machine learning. 

• Détail de l'offre :  https://www.irit.fr/SAMOVA/site/wp-content/uploads/2021/12/ALAIA_Ingenieur_de_Recherche_2022_Franc%CC%A7ais.pdf
  
• Lien vers le LabCom : https://anr.fr/Projet-ANR-18-LCV3-0001 et https://www.irit.fr/SAMOVA/site/projects/current/labcom-alaia/
• Salaire : selon la grille en vigueur, le profil et l'expérience
• Date limite de candidature : les deux postes sont à pourvoir au plus tôt,
  
• Prise de poste : début février 2022 si possible
  

Les candidatures sont à adresser à Isabelle Ferrané (isabelle.ferrane@irit.fr) et Lionel Fontan (lfontan@archean.tech). N'hésitez pas à nous contacter pour de plus amples informations. 

Top

1) Research Associate in Safe Conversational AI(re-advertising)

Top

Aalto opens a call for an assistant or associate professor in speech and language technology. 

https://aalto.wd3.myworkdayjobs.com/PrivateJobPosting/job/Otaniemi-Espoo-Finland/ASSISTANT-OR-ASSOCIATE-PROFESSOR-IN-SPEECH-AND-LANGUAGE-TECHNOLOGY--tenure-track-_R32437

Top

We are looking to recruit two Post-Doctoral Researchers to join the UCD team in the ADAPT Research Centre (www.adaptcentre.ie).  These projects are part of the ADAPT Digital Content Transformation Strand.

 

1) Harnessing Speech for Social Inclusion

Rather than focusing on accuracy and error rates, evaluation of speech recognition systems should be contextualized with respect to how well they perform in situations where the interlocutor is not the ‘typical’ native speaker (e.g. a senior citizen, a citizen with disabilities or a non-native speaker). Through publicly engaged research, data will be collected on how well existing speech technology is serving our citizens. A systematic evaluation of the output data produced by existing ASR systems and the interactions that arise when an ASR ‘converses’ with a range of users will enable a categorization of interaction issues and error patterns that need to be accounted for when developing applications which provide interfaces to essential services. This project will involve the development of a system which facilitates diagnostic evaluation of ASRs in a variety of interaction scenarios providing linguistic cues the augmentation of the ASR and building on the low-resource MT technologies being developed in other ADAPT projects.

 

2) Embedding of Multi-level Speech Representations

While deep learning has led to huge performance gains in speech recognition and synthesis, only recently more focus is being placed on what deep learning may be able uncover about the patterns which humans use intuitively when interacting via speech and which distinguish native from non-native speakers. Such patterns are typically the focus of speech perception and experimental phonetic studies. This project aims to build on the notion of multi-linear or multi-tiered representations of speech, creating embeddings of multiple (sub- word) levels of representation – phonetic features, phonemes, syllable pieces and syllables – enabling a closer investigation of systematicity and variability of speech patterns. This research will find application in non-native speech recognition, in speech adaptation/accommodation for native and non-native interactions and in pronunciation training scenarios.

 

ADAPT is the world-leading SFI research centre for AI Driven Digital Content Technology, coordinated by Trinity College Dublin and based within Dublin City University, University College Dublin, Technological University Dublin, Maynooth University, Munster Technological University, Athlone Institute of Technology, and the National University of Ireland Galway. ADAPT’s research vision is to pioneer new forms of proactive, scalable, and integrated AI-driven Digital Content Technology that empower individuals and society to engage in digital experiences with control, inclusion, and accountability with the long-term goal of a balanced digital society by 2030. ADAPT is pioneering new Human Centric AI techniques and technologies including personalisation, natural language processing, data analytics, intelligent machine translation human-computer interaction, as well as setting the standards for data governance, privacy and ethics for digital content.

ADAPT Digital Content Transformation Strand
From the algorithmic perspective, new machine learning techniques will both enable more users to engage meaningfully with the increasing volumes of content globally in a more measurably effective manner, while ensuring the widest linguistic and cultural inclusion. It will enhance effective, robust integrated machine learning algorithms needed to provide multimodal content experiences with new levels of accuracy, multilingualism and explainability.

 

Full details of the positions, requirements and link to submit applications can be found at:

 

1) https://www.adaptcentre.ie/careers/postdoctoral-researcher-in-harnessing-speech-for-social-inclusion/

2) https://www.adaptcentre.ie/careers/postdoctoral-researcher-in-embedding-multi-level-speech-representations-2/

 

The closing date is 17:00hrs (local Irish time) on Monday 14th February 2022 and and candidates must apply via https://www.ucd.ie/workatucd/jobs/  

 

The reference is 

 

1) 014028 for Harnessing Speech for Social Inclusion

2) 014079 for Embedding of Multi-level Speech Representations

 

Top

A public selection procedure is called, based on qualifications and interview, for the assignment of n. 2 research fellowship grant for collaboration in research activities

Project main aim: Multidisciplinary Research on AI for Health.

Location: Kore University of Enna - Enna (Sicily), Italy

Funding Programme: Research Projects of National Relevance - PRIN 2017

Description: The project focuses on idiopathic Parkinson’s disease dysarthric speech, produced by speakers of two varieties of Italian that show different segmental (consonantal, vocalic) and prosodic characteristics. The project as a whole aims at: identifying phonetic features that impact on speech intelligibility and accuracy, separating variability due to dysarthria from features due to sociolinguistic variation, and developing perspectives and tools for clinical practice that take variation into account.

Duration: 12 months

Link to apply: https://unikore.it/index.php/it/contratti-di-ricerca/item/41282-d-p-n-33-2022-2-assegni-di-ricerca-presso-l-universita-degli-studi-di-enna-kore

Further information contact Prof. Sabato Marco Siniscalchi, E-mail: marco.siniscalchi-at-unikore.it

Top

Job description

One of the most pressing matters that holds back robots from taking on more tasks and reach a widespread deployment in society is their limited ability to understand human communication and take situation-appropriate actions. This PhD positions is dedicated to addressing this gap by developing the underlying data-driven models that enable a robot to engage with humans in a socially aware manner.

This position is specifically targeted at the development of an argumentative dialogue system for human-robot interaction.  The PhD candidate will explore how to fuse multimodal behaviour to infer a person's perspective. The candidate will use, and further develop, reinforcement learning techniques in order to drive the robot's argumentative strategy for deliberating topics of current social importance such as global warming or vaccination.

The ideal candidate will have a keen interest in speech technology and reinforcement learning. He or she has strong interactive system background will design and run the experiments to evaluate the created hybrid-AI models through human-robot interaction.

Topics of interest:

1) long-term human-robot interaction

2) affective computing

2) NLP&argument-mining

Requirements

• MSc in Computer Science or related field.
• At least 3 years of programming experience in python (java  or C++is a plus).
• Motivation to meet deadlines.
• Affinitty to design and social science research.
• Interest in collaborating with colleagues from Industrial Design.
• Willingness to teach and guide students.
• The ability to work in a team, take initiative, be results oriented and systematic.

Top

Postdoctoral position in Speech Processing at INRIA, Bordeaux, France

Title: Glottal source inverse filtering for the analysis and classification of pathological speech

Keywords: Pathological speech processing, Glottal source estimation, Inverse filtering, Machine learning, Parkinsonian disorders, Respiratory diseases

Contact and Supervisor: Khalid Daoudi (khalid.daoudi@inria.fr)

INRIA team: GEOSTAT (geostat.bordeaux.inria.fr)

Duration: 13 months (could be extended)

Starting date: between 01/04/2022 and 01/06/2022 (depending on the candidate availability)

Application : via https://recrutement.inria.fr/public/classic/en/offres/2022-04481

Salary: 2653€/month (before taxes, net salary 2132€)

Profile: PhD thesis in signal/speech processing (or a solid post-thesis experience in the field) Required Knowledge and background: A solid knowledge in speech/signal processing; Basics of machine learning; Programming in Matlab and Python. Scientific research context During this century, there has been an ever increasing interest in the development of objective vocal biomarkers to assist in diagnosis and monitoring of neurodegenerative diseases and, recently, respiratory diseases because of the Covid-19 pandemic. The literature is now relatively rich in methods for objective analysis of dysarthria, a class of motor speech disorders [1], where most of the effort has been made on speech impaired by Parkinson’s disease. However, relatively few studies have addressed the challenging problem of discrimination between subgroups of Parkinsonian disorders which share similar clinical symptoms, particularly is early disease stages [2]. As for the analysis of speech impaired by respiratory diseases, the field is relatively new (with existing developments in very specialized areas) but is taking a great attention since the beginning of the pandemic. The speech production mechanism is essentially governed by five subsystems: respiratory, phonatory, articulatory, nasalic and prosodic. In the framework of pathological speech, the phonatory subsystem is the most studied one, usually using sustained phonation (prolonged vowels). Phonatory measurements are generally based on perturbations or/and cepstral features. Though these features are widely used and accepted, they are limited by the fact that the produced speech can be a product of some or all the other subsystems. The latter thus all contribute to the phonatory performance. An appealing way to bi-pass this problem is to try to extract the glottal source from speech in order to isolate the phonatory contribution. This framework is known as glottal source inverse filtering (GSIF) [3]. The primary objective of this proposal is to investigate GSIF methods in pathological speech impaired by dysarthria and respiratory deficit. The second objective is to use the resulting glottal parameterizations as inputs to basic machine learning algorithms in order to assist in the discrimination between subgroups of Parkinsonian disorders (Parkinson’s disease, Multiple-System Atrophy, Progressive Supranuclear Palsy) and in the monitoring of respiratory diseases (Covid-19, Asthma, COPD). Both objectives benefit from a rich dataset of speech and other biosignals recently collected in the framework of two clinical studies in partnership with university hospitals in Bordeaux and Toulouse (for Parkinsonian disorders) and in Paris (for respiratory diseases).

Work description GSIF consists in building a model to filter out the effect of the vocal tract and lips radiation from the recorded speech signal. This difficult problem, even in the case of healthy speech, becomes more challenging in the case of pathological speech. We will first investigate time-domain methods for the parameterization of the glottal excitation using glottal opening and closure instants. This implies the development of a robust technique to estimate these critical time-instants from dysarthric speech. We will then explore the alternative approach of learning a parametric model of the entire glottal flow. Finally, we will investigate frequency-domain methods to determine relationships between different spectral measures and the glottal source. These algorithmic developments will be evaluated and validated using a rich set of biosignals obtained from patients with Parkinsonian disorders and from healthy controls. The biosignals are electroglottography and aerodynamic measurements of oral and nasal airflow as well as intra-oral and sub-glottic pressure. After dysarthric speech GIFS analysis, we will study the adaptation/generalization to speech impaired by respiratory deficits. The developments will be evaluated using manual annotations, by an expert phonetician, of speech signals obtained from patients with respiratory deficit and from healthy controls. The second aspect of the work consists in manipulating machine learning algorithms (LDA, logistic regression, decision trees, SVM…) using standard tools (such as Scikit-Learn). The goal here will be to study the discriminative power of the resulting speech features/measures and their complementarity with other features related to different speech subsystems. The ultimate goal being to conceive robust algorithms to assist, first, in the discrimination between Parkinsonian disorders and, second, in the monitoring of respiratory deficit.

Work synergy - The postdoc will interact closely with an engineer who is developing an open-source software architecture dedicated to pathological speech processing. The validated algorithms will be implemented in this architecture by the engineer, under the co-supervision of the postdoc. - Giving the multidisciplinary nature of the proposal, the postdoc will interact with the clinicians participating in the two clinical studies.

References: [1] J. Duffy. Motor Speech Disorders Substrates, Differential Diagnosis, and Management. Elsevier, 2013. [2] J. Rusz et al. Speech disorders reflect differing pathophysiology in Parkinson's disease, progressive supranuclear palsy and multiple system atrophy. Journal of Neurology, 262(4), 2015. [3] P. Alku. Glottal inverse filtering analysis of human voice production – A review of estimation and parameterization methods of the glottal excitation and their applications. Sadhana – Academy Proceedings in Engineering Sciences. Vol. 36, Part 5, pp. 623-650, 2011

Top

La jeune compagnie EasyChain et le laboratoire de recherche L3i de La Rochelle lancent un appel à candidatures pour un poste de doctorant.e CIFRE dans le domaine de développement des agents conversationnels.

Les détails de l'offre sont disponibles à cette adresse :
Dynamic Human-Agent Interactions adapted to users? profiles.

Le.la candidat.e retenu.e devra être titulaire d'un master ou d'un diplôme équivalent en informatique ou en traitement automatique des langues. Un solide bagage en apprentissage automatique et une bonne communication en anglais sont requis.

La thèse se déroulera principalement à Niort chez EasyChain, dans un environnement francophone.

Si vous êtes intéressé.e par ce poste, veuillez envoyer les informations suivantes à Antoine Doucet (antoine.doucet@univ-lr.fr),  et Ahmed Hamdi (ahmed.hamdi@univ-lr.fr):
 * CV détaillé
 * Diplômes de licence et de master.
* Lettres de recommandation

Les candidatures seront étudiées jusqu'au 1 mars 2022.

Bien cordialement

Antoine Doucet

Top