1 post-doc in Brain- and biosignal-based speech recognition

https://beta.iit.it/careers/openings/opening/66-postdoctoral-position-in-brain-and-biosignal-based-speech-recognition

- 1 post-doc in New techniques for vision-assisted speech processing with event-drive cameras (within the EU H2020 Ecomode project) 

https://beta.iit.it/careers/openings/opening/71-postdoctoral-position-in-new-techniques-for-vision-assisted-speech-processing

Back

Top

Un poste de Maître de Conférences en informatique (section 27) est mis au concours pour
la campagne 2016 au Centre d'Enseignement et de Recherche en Informatique de l'Université
d'Avignon (ceri.univ-avignon.fr).

La personne recrutée effectuera sa recherche au sein du LIA (lia.univ-avignon.fr).

Le profil du poste est consultable sur lia.univ-avignon.fr rubrique Emplois et sur
GALAXIE :

- Informatique - Profil indexation audio/biométrie vocale  (Poste 088 / Galaxie 4080) -
https://www.galaxie.enseignementsup-recherche.gouv.fr/ensup/ListesPostesPublies/ANTEE/2016_1/0840685N/FOPC_0840685N_4080.pdf

Contacts :

Enseignement : Corinne Fredouille corinne.fredouille@univ-avignon.fr - Directrice des
Etudes du CERI et Fabrice Lefèvre fabrice.lefevre@univ-avignon.fr - Directeur du CERI

Recherche : Fabrice Lefèvre fabrice.lefevre@univ-avignon.fr - Directeur du CERI,
Directeur LIA par interim

Back

Top

Trois postes d'enseignants-chercheurs en informatique sont à pourvoir à l'Université du Maine lors de la campagne de recrutement 2016 : un poste de Professeur et deux postes de Maîtres de conférences.

Les personnes recrutées sur chacun de ses postes effectueront leurs travaux de recherche au LIUM (Laboratoire d'Informatique de l'Université du Maine) :

- Le poste de Professeur est rattaché au Département d'informatique de  l'UFR Sciences de l'Université du Maine. Le profil recherche est spécialisé dans le domaine du traitement automatique du locuteur. Le profil enseignement  vise les matières fondamentales de l'informatique, incluant la capacité à enseigner la théorie des langages.
Plus d'info :
https://www.galaxie.enseignementsup-recherche.gouv.fr/ensup/ListesPostesPublies/ANTEE/2016_1/0720916E/FOPC_0720916E_4154.pdf

- Un poste de Maître de conférences est rattaché au Département d'informatique de l'UFR Sciences de l'Université du Maine. Le profil recherche est spécialisé dans le domaine de la reconnaissance automatique de la parole. Le profil enseignement  vise les matières fondamentales de l'informatique, incluant la capacité à enseigner la théorie des langages.
Plus d'info :
https://www.galaxie.enseignementsup-recherche.gouv.fr/ensup/ListesPostesPublies/ANTEE/2016_1/0720916E/FOPC_0720916E_4153.pdf

- L'autre poste de Maître de conférences est rattaché au Département Gestion des Entreprises et des Administrations de l'IUT du Mans. Le profil recherche est spécialisé dans le domaine de l'apprentissage automatique et de la fouille de données textuelles. Le profil enseignement concerne l'informatique de gestion, les bases de données et l'informatique appliquée à la gestion des entreprises.
Plus d'info :
https://www.galaxie.enseignementsup-recherche.gouv.fr/ensup/ListesPostesPublies/ANTEE/2016_1/0720083Z/FOPC_0720083Z_4170.pdf

Back

Top

Vous trouverez à l'URL suivante et en pièce jointe une offre de thèse ministérielle à l'IRISA intitulée 'Transformation de registres par extraction de motifs langagiers' : https://www-expression.irisa.fr/files/2011/07/sujet1.pdf . Merci de diffuser auprès de personnes éventuellement intéressées.

Pour plus de précisions ou pour postuler, merci de contacter :

- Pascale Sébillot : pascale.sebillot@irisa.fr

- Nicolas Béchet : nicolas.bechet@irisa.fr

- Gwénolé Lecorvé : gwenole.lecorve@irisa.fr

Date limite de candidature : 30 avril 2016.

Back

Top

We are happy to advertise for a Post Doctoral Position in Machine Learning at Univ. Lille, France

Topic
Machine Learning for Human-Robot Interaction

Job description
This position is offered in the framework of the BabyRobot H2020 (www.babyrobot.eu) project involving 8 partners across Europe (including France, Sweden, UK, Greece, Germany and Denmark). The project focuses on interactive robotics and especially on interaction with children. 

Breakthroughs in core robotic technologies are needed to support this research mainly in the areas of motion planning and control in constrained spaces, gestural kinematics, sensorimotor learning and adaptation. In addition, new models of interaction need to be developed. Because of the human being in the loop, standard control theory can hardly be applied. For this reason, machine learning methods such as reinforcement and imitation learning have been identified as candidates to address these issues in a unied framework. Therefore, the applicant will be involved in research in core machine learning applied to control and interaction. Several directions of research are envisioned. First, recent works on stochastic games [3] applied to dialogue management [1] can be further investigated so as to be adapted to multimodal and multiparty interaction scenarios. Second, the 'learning from demonstration' (LfD) [4] paradigm can be adapted to the adversarial case so as to transfer interactional behaviours from actual human-human interactions to machines. Other  topics can be investigated such as inverse reinforcement learning [2] or transfer learning.

Profile
The applicant should have completed a PhD in computer science, statistical learning or robotics. The ideal candidate will have a strong background in machine learning and especially in reinforcement learning or stochastic games. Experience in interactive systems (spoken dialogue systems, interactive robotics, human-machine interfaces) would be much appreciated. The recruited person will be involved in the management of the project, participate to consortium meetings and contribute to deliverables. Therefore, good communication skills and autonomy are mandatory. Preference will go to candidate with a strong publication record.

Work environment
The position is offered in the Sequential Learning (SequeL) research team (joint team between Inria, Univ. Lille and CNRS) located in Lille, France. SequeL is a world-leading group in reinforcement learning, bandit theory and recommendation systems involving 30 members (including 10 permanent staff members). The team's working language is English. The team is part of the French National Institute for Computer Science and Mathematics (Inria) as well as the Computer Science and Signal Processing laboratory of Lille (CRIStAL). Lille is the capital of the north of France, a metropolis with 1 million inhabitants, with excellent train connection to Brussels (30 min), Paris (1h) and London (1h30).

How to apply
The application should include a brief description of research interests and past experience, a CV, degrees and grades, motivation letter, relevant publications, letter(s) of recommendation and contact information to reference persons.

Details
Application deadline: 30th of April 2016
Starting date: May or June 2016
Duration: 24 months (can be extended)
Salary (after taxes): 2100 euros

Contacts
Olivier Pietquin: olivier.pietquin@univ-lille1.fr
Bilal Piot: bilal.piot@univ-lille1.fr

References
[1] Merwan Barlier, Julien Perolat, Romain Laroche, and Olivier Pietquin. Human-machine dialogue as a stochastic game. In Proceedings of the 16th Annual SIGdial Meeting on Discourse and Dialogue (SIGDIAL 2015), pages 2{11, Prague (Czech Republic), September 2015.
[2] Edouard Klein, Matthieu Geist, Bilal PIOT, and Olivier Pietquin. Inverse Reinforcement Learning through Structured Classication. In Advances in Neural Information Processing Systems (NIPS 2012), pages 1007{1015, Lake Tahoe (NV, USA), December 2012.
[3] Julien Perolat, Bruno Scherrer, Bilal Piot, and Olivier Pietquin. Approximate dynamic programming for two-player zero-sum markov games. In Proceedings of the International Conference on Machine Learning (ICML 2015), Lille (France), July 2015.
[4] Bilal Piot, Matthieu Geist, and Olivier Pietquin. Learning from demonstrations: Is it worth estimating a reward function? In Hendrik Blockeel, Kristian Kersting, Siegfried Nijssen, and Filip Zelezny, editors, Proceedings of the European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML/PKDD 2013), volume 8188 of Lecture Notes in Computer Science, pages 17{32, Prague (Czech Republic), September 2013. Springer.

Back

Top

Research Associate Position at York University, Toronto, Canada

The Lassonde School of Engineering, York University (http://lassonde.yorku.ca/) is sponsoring a fulltime
research associate position in deep learning for natural language processing (NLP). This position
is comparable with a research professorship (with annual salary up to $85k), an initial two-year
appointment but extendable. This research position is created to support the cutting-edge research
activities in NLP, particularly applying deep learning to the challenging NLP or AI problems, including
(not limited to) information extraction, language understanding and inference, and probabilistic
reasoning.
If you are interested, please email your CV to:
Prof. Hui Jiang (hj@cse.yorku.ca) (http://wiki.eecs.yorku.ca/user/hj/)
iFLYTEK Laboratory for Neural Computing and Machine Learning (iNCML)
(http://wiki.eecs.yorku.ca/lab/MLL/start)
Department of Electrical Engineering and Computer Science
The Lassonde School of Engineering, York University
The deadline is May 15, 2016.

Back

Top

A new PhD position is open in 'Multilingual Affective Speech Synthesis' at Idiap Research
Institute:
 http://www.idiap.ch/education-and-jobs/job-10193

Candidates with knowledge of French and English would be at an advantage.

Back

Top

EMR.AI Inc., San Francisco, CA

RESEARCH SCIENTIST, NATURAL LANGUAGE PROCESSING

Headquartered in San Francisco, CA, EMR.AI Inc. is a leading provider of AI solutions to the medical sector. EMR.AI transforms unstructured information, in form of written, spoken, or typed reports, clinical test results, and radiographs into international standard codes saved in common EMR systems. The wealth of discrete medical data provided through this transformation in conjunction with EMR.AI's suite of medical analytics solutions enables stakeholders, practitioners, researchers, health providers, and policy makers to obtain a comprehensive picture of the available medical data in their organization.

SUMMARY

EMR.AI Research & Development has openings for Research Scientists in the field of Natural Language Processing in our Downtown San Francisco offices. Scientists will work on projects spanning a variety of tasks including the semantic interpretation of written and spoken medical reports, the design of language models for a variety of NLP tasks and speech recognition, the summarization of written and spoken language in the medical domain, the incorporation of lexica, ontologies, relational databases, and other sources of structured and unstructured knowledge sources into EMR.AI’s medical NLP tool set, and others. This is a unique opportunity to be part of a cutting-edge R&D team in the epicenter of the world’s AI tech industry with true impact on medical research.

RESPONSIBILITIES

 Process huge corpora of medical textual documents to perform syntactic and semantic analyses and train, tune, and test probabilistic and other data-driven models, using both existing tool benches, proprietary and open-source, as well as self-developed algorithms and techniques.

 Produce high-quality programs and scripts to embed scientific algorithms into effective prototypes and demos to be shared with EMR.AI’s leadership team, its customers, partners, and vendors.

 Create and document technological innovations by means of patent disclosures, scientific publications, media alerts, and other channels.

 Work closely with EMR.AI’s speech processing team and its software engineering division to produce innovative and effective solutions for a range of AI products and services in the medical domain.

 Represent the R&D division in communications with EMR.AI’s leadership team, its customers, partners, and vendors at meetings, conventions, and other venues as well as in written statements.

SKILLS

PhD in computer science, computational linguistics, electrical engineering, or a related field. Experience in the state of the art of NLP and its standard tools is required. Candidates must be very skilled in programming and must have a proven scientific track record. They must be excellent team players, including with distributed teams, and strong in oral and written English communication. Knowledge of the US medical sector is desirable, so are experience with start-ups and strong scientific connections throughout the Bay Area and beyond.

BENEFITS

EMR.AI offers competitive salaries, an excellent benefit package, and a stimulating work environment in the heart of San Francisco with manifold local, domestic, and international commercial and academic partnerships.

HOW TO APPLY

Please send your application documents to jobs@emr.ai

CONTACT

EMR.AI Inc. 90 New Montgomery St San Francisco, CA 94105,

USA phone: +1-415-200-8535

e-mail: info@emr.ai www: http://emr.ai

Back

Top

A 3-year postdoctoral research position in speech processing is available in the Sigmedia Group in Trinity College Dublin, Ireland.

Full details of the position are given in the advert available at: http://adaptcentre.ie/careers/TCD_postdoc_themeD-2.pdf

Back

Top

Postdoctoral Position in Linguistics (Prosodic Phrasing and Discourse Comprehension), Université catholique de Louvain

The Institute for Language & Communication of the Université catholique de Louvain (Belgium) is currently seeking a postdoctoral researcher to perform corpus analysis and experimental research on the role of prosody in discourse comprehension. The successful candidate will work under the supervision of Prof. Anne Catherine Simon and will join the project “PhraDiCo: Prosodic phrasing for discourse comprehension. Perception of prosodic boundaries and their contribution to the construal of discourse coherence”. He/she will be employed on a fixed-term contract for 24 months.

Project Description

[PhraDiCo] Prosodic phrasing for discourse comprehension. Perception of prosodic boundaries and their contribution to the construal of discourse coherence. Prosodic phrasing is central to language comprehension and discourse coherence. Spoken language is highly transitory and dynamic. Speech is processed in real-time, as the hearer allocates his attention to each incoming unit and integrates it into a coherent whole. Speech is chunked into prosodic units which are presumably used as basic units of discourse processing. During its first phase, this project aims to explore how ordinary listeners (as opposed to trained experts) perceive prosodic boundaries in spoken language. What is the relationship between the prosodic cues (pauses, lengthening, melodic discontinuity etc.) and the perceived strength of prosodic boundaries? To what extent does the syntactic structure contribute to the perception of prosodic boundaries? Are naïve listeners consistent in their appreciation of boundary strength on word boundaries? Once those questions have been sufficiently tackled, we will focus on analysing the contribution of prosodic boundaries to discourse coherence: relational coherence, which seeks to establish discourse relations between discourse “building blocks”, and referential coherence, i.e. the identification of large-scale discourse topics. Four perceptual experiments have been designed to investigate (i) the acoustic and linguistic factors involved in the perception of prosodic boundary strength; (ii) the influence of speaking styles variability (speech rate, pitch register) on prosodic boundary perception; (iii) the contribution of prosodic units to the construal of discourse relations and (iv) to the identification of large-scale discourse topics.

The project is undertaken by the research centre “Valibel – Discours & Variation” (Discourse & Variation, www.uclouvain.be/valibel) and is funded by F.S.R. – FNRS (www.frs-fnrs.be).

Job Description

• Discursive and prosodic analysis of speech data (transcription, alignment and annotation).

• Design of experimental procedures and preparation of speech samples for perceptual experiments.

• Organisation of on-line and off-line speech perception experiments (recruitment of participants, technical and practical aspects of execution etc.).

• (Statistical) Analysis of the experimental results using databases and specialised tools.

The appointed researcher will work in close collaboration with the project’s supervisor and another postdoctoral researcher. He/she will participate in research leading to scientific publications.

Qualifications

• PhD in Linguistics (discourse analysis, experimental linguistics, cognitive approach to discourse comprehension, syntax-prosody interface) obtained not more than 6 years before the starting date.

• Experience in discourse annotation (topic-focus, discourse relations), in experimental research (perceptual studies), and/or in prosodic analysis.

• Expertise in statistical methods for analysing linguistic data is desirable.

• Knowledge of the Praat software is an asset.

• Native-like command of French and good command of academic English.

Place of employment: Louvain-la-Neuve, Belgium (30 km south of Brussels). With 21000 students and 5000-member staff, the University of Louvain is the largest university of the Brussels and Wallonia regions of Belgium.

Contract length: 24 months.

Contract term: Full-time

Starting date: July 2016

Deadline for applications: 5 June 2016

Applications (motivation letter, detailed CV, 1-3 recommendation letters and 3 relevant publications), should be sent in electronic format to Prof. Anne Catherine Simon (anne-catherine.simon@uclouvain.be).

Back

Top

PhD positions are available in our group at Inria, in France, on deep learning for:
- speech enhancement
- robust speaker recognition
- robust speech recognition
- language model adaptation

Details:
http://www.adum.fr/as/ed/voirproposition.pl?langue=en&site=IAEM&matricule_prop=12468
http://www.adum.fr/as/ed/voirproposition.pl?langue=en&site=IAEM&matricule_prop=12465
http://www.adum.fr/as/ed/voirproposition.pl?langue=en&site=IAEM&matricule_prop=12473
http://www.adum.fr/as/ed/voirproposition.pl?langue=en&site=IAEM&matricule_prop=10773

Profile:
Candidates must hold an MSc in computer science, machine learning, signal processing, or
applied mathematics, or be about to obtain one. Programming experience in Python or C/C++
is a must.

Working Environment:
Nancy is one of the top cities to study in France with 50,000 students, cheap
accomodation, a vibrant cultural scene, and good connections to Paris (1.5h), Luxemburg
(1.5h), Belgium, and Germany. Inria Nancy is a 500-people research institute dedicated to
computer science. The Multispeech team (https://team.inria.fr/multispeech/) is a
30-people group covering various fields of speech science, with a strong emphasis on
statistical machine learning and signal processing.

To apply:
Click on 'candidater' at the bottom of the webpage. You will be asked for a detailed CV,
BSc and Msc transcripts and diplomas, publications, and recommendation letters. Please
apply as soon as possible. There will be a first cutoff date on May 31 and a second one
on June 30.

--
Emmanuel Vincent

Back

Top

PhD Position: Merging acquisition and processing of cineMRI of the vocal tract

Subject

Tracking the position of speech articulators along time is crucial to understand speech production better. For a long time X-ray imaging was the only technology able to acquire images at a sufficiently high sampling frequency (around 50 images per second) for visualizing articulatory gestures. However, this technique has been abandoned at the end of the eighties because of the health hazard implied by ionizing radiations. Furthermore, since the whole vocal tract is projected on the image plane contours of organs (especially the mandible, teeth and tongue) overlap on the images making the processing of images very difficult.

The interest of Magnetic Resonance Imaging (MRI) is to provide an excellent contrast of soft tissues for a slice placed in any orientation and dynamic MRI is acknowledged as a powerful tool for imaging speech production [3].  However, current performance of cineMRI remains inadequate in terms of sampling rate and spatial resolution, and the objective of this thesis is to develop more efficient acquisition protocols and algorithms.

The objective is to develop protocols by exploiting the latest advances in MRI, particularly parallel imaging and reconstruction under parsimony constraints called ?compressed sensing? [1]. IADI laboratory developed MRI reconstruction techniques with movement compensation [2] and multi-slice dynamic reconstruction enabling super resolution. These techniques have been already applied to cardio-respiratory movements by using physiological signals (ECG and respiratory) as constraints for the reconstruction algorithms.

A first preliminary work consisting of applying these techniques to the domain of speech production  has been carried out [4]. A second work was dedicated to the development of an acquisition protocol based on ?compressed sensing?. The idea is to exploit the parsimony of the image Fourier transform coefficients in order to acquire only a small number of them, and then to reconstruct the image in an optimal manner.

However, it is possible to do better since the speech signal is acquired simultaneously then denoised before being segmented into speech sounds. Therefore, the contribution of each line acquired in the image Fourier space can be related to the speech sound it corresponds to, and one can take advantage of this information to improve the resolution of reconstructed images. This  idea will be exploited with the objective of realizing a proof of concept of automatic acquisition/reconstruction of MRI images of vocal tract during speech production. We would like to go further by utilizing the knowledge of the speech sound and the approximative vocal tract shape predicted for the sound by an articulatory model to pilot acquisition.

Keywords: MRI acquisition, compressed sensing, speech processing, articulatory modeling, vocal tract

Some references

[1]Michael Lustig, David Donoho, and John M. Pauly. Sparse MRI: The application of compressed sensing for rapid MR imaging. MAGNETIC RESONANCE IN MEDICINE, 58(6):1182?1195, December 2007.

[2]F. Odille, P. A. Vuissoz, P. Y. Marie, and J. Felblinger. Generalized reconstruction by inversion of coupled systems (GRICS) applied to free-breathing MRI. Magn Reson Med, 60(1):146?57, July 2008.

[3]Andrew D Scott, Marzena Wylezinska, Malcolm J Birch, and Marc E Miquel. Speech MRI: Morphology and function. Physica Medica, 30(6):604?618, 2014.

[4]P.A. Vuissoz, F. Odille, Y. Laprie, E. Vincent, G. Hossu, and J. Felblinger. Speech Cine SSFP with optical microphone synchronization and motion compensated reconstruction. In ISMRM Workshop on Motion Correction in MRI, Tromso, Norvège, May 2014.

Environment

Both laboratories IADI and LORIA have developed a narrow and fruitful collaboration for years which in particular resulted in the development of a ?compress sensing? acquisition algorithm and in a research contract on articulatory synthesis. A working environment covering articulatory modeling and MRI data acquisition domains is now available and will offer very favorable conditions for this work.

Supervisors

·       Pierre-André Vuissoz (IADI ? Imagerie Adaptative Diagnostique et Interventionnelle, unité INSERM U947) pa.vuissoz@chu-nancy.fr

·       Yves Laprie (MultiSpeech team at LORIA ? Laboratoire Lorrain de Recherche en Informatique et ses Applications, UMR 7503)

Application

We are looking for a highly motivated person with a master degree in computer sciences, applied mathematics or computer sciences. The applicant should have a solid background in signal processing (and Matlab software) and computer sciences. Knowledge in speech processing will be also appreciated.

Interested applicants should send their application file to Yves.Laprie@loria.fr and pa.vuissoz@chru-nancy.fr:

a)     Motivation letter,

b)     CV,

c)     academic transcripts (with explanation of the grade scale adopted),

d)     2 references (letters or names)

Expected start date: 1st October 2016

Back

Top

Title: Senior Speech Analytics Engineer

Job Description

SRI’s Speech Technology and Research (STAR) Lab seeks a self-motivated, team-oriented, experienced leader and innovator to work in the area of speaker state analysis (for example, emotion and health states). The work will include advancing technology capabilities, determining client needs and understanding client data, winning new funding, interacting with clients, both government and commercial, and leading current or new projects in this area. The successful candidate must be proactive and entrepreneurial. This is an opportunity to have a significant impact in an emerging research area.

STAR Lab engages in leading-edge research in speech recognition, speaker characterization, speaker and language identification, machine translation, natural language processing and other areas of speech/language technology, offering opportunities from basic research to prototyping, productization, and widespread deployment. The multidisciplinary research team consists of excellent speech researchers, linguists and software engineers. Characteristics of STAR staff are enthusiasm, self-motivation, initiative, passion for learning, taking ownership, thriving in a flat organizational hierarchy, and a desire to control their own career paths.

SRI International is a premier nonprofit research organization based in Menlo Park, California with a mission to create world-changing solutions making people safer, healthier, and more productive.

Located in the heart of Silicon Valley, the world’s center for innovation and technology, SRI’s focus spans everything from basic research to delivered systems, government and commercial work and spinning off successful companies like SIRI. Flexible cross-laboratory teams often form dynamically to solve challenging problems spanning multiple disciplines.

Requirements

A PhD degree or equivalent with at least 10 years of experience in speech or NLP technology. Prior interest in the area of speaker characterization is an advantage. Experience of leading teams of projects in either academia or industry along with a record of either successful proposals (government or industry) or managing a start-up in the speech area is required. Strong oral and written communication skills in English are expected. US citizenship is a plus but not required. Must be flexible to traveling.

Apply online at www.sri.com/careers

Job ID: 1020BR

SRI is an Equal Employment Opportunity/Affirmative Action Employer

Back

Top

Research Associate Position at York University, Toronto, Canada

The Lassonde School of Engineering, York University (http://lassonde.yorku.ca/) is
sponsoring a fulltime
research associate position in deep learning for natural language processing (NLP). This
position
is comparable with a research professorship (with annual salary up to $85k), an initial
two-year
appointment but extendable. This research position is created to support the cutting-edge
research
activities in NLP, particularly applying deep learning to the challenging NLP or AI
problems, including
(not limited to) information extraction, language understanding and inference, and
probabilistic
reasoning.
If you are interested, please email your CV to:
Prof. Hui Jiang (hj@cse.yorku.ca) (http://wiki.eecs.yorku.ca/user/hj/)
iFLYTEK Laboratory for Neural Computing and Machine Learning (iNCML)
(http://wiki.eecs.yorku.ca/lab/MLL/start)
Department of Electrical Engineering and Computer Science
The Lassonde School of Engineering, York University
 

Back

Top

Speech Technologist @ Voice Tech Startup!

Location: Belmont, MA

Available: Immediately

VocaliD is a voice technology start-­up that is bringing speaking machines to life by creating

personalized digital voices. Our award-­winning technology is grounded in over a decade of

research at Northeastern University and has enabled us to attract Small Business Innovation

awards from the National Science Foundation and the National Institutes of Health. We

successfully delivered a set of beta voices in 2015 and have a waiting list of pre-­orders

though our oversubscribed crowdfunding campaign. We are now seeking an experienced

research-­based speech technologist to join our team as we launch our personalized voice

service in mid 2016.

Communications applications and devices are getting smarter and using more sophisticated

AI, yet the voices remain robotic and unnatural. Until now, the cost and time to build a custom

voice has been prohibitive. VocaliD creates personalized digital voices at scale for use on

any device that speaks. Our proprietary voice search and speech blending algorithms

leverage our “Human Voicebank”, a collection of 3.5 million of recordings crowdsourced from

around the world. Our breakthrough is that we only need 2-­3 seconds of someone’s voice to

(re)create a voice when combined with recordings of a matched speaker from the Voicebank.

Our initial market addresses the 10M+ people worldwide who use a machine to speak.

VocaliD enables these individuals to speak in their own voice instead of the familiar “Stephen

Hawking” voice. Join our dynamic team as we push the envelop on innovation while having

real social impact!

Responsibilities:

-­ Create tools for filtering and parsing recorded voices

-­ Implement machine learning algorithms to match donor and recipient voices

-­ Enhance source-­filter separation and voice conversion algorithms

-­ Design and implement novel algorithms for tracking and monitoring speech changes

-­ Create a framework for algorithmic synthesis of voices from donor samples

-­ Implement a SaaS framework for voice deployment

Required Qualifications:

-­ MS or PhD in Electrical or Computer Engineering, Computer Science or related field

-­ Experience programming in C/C++, Python, Java

-­ Experience productizing algorithms from prototypes in Matlab (or other)

-­ Experience with Digital Signal Processing and ideally speech/voice technologies

-­ Experience with AWS services EC2, S3, RDS (Postgres), Route 53

-­ Experience with Windows and Linux

If you are interested in joining a thriving and passionate team, please submit a cover letter

and resume to rupal@vocaliD.co

Visit us at www.vocaliD.co for more information about VocaliD.

Back

Top

Call for Postdoctoral Positions in Cognitive Science/Psychology/Linguistics

Institution: Laboratoire Parole et Langage (CNRS, Aix-Marseille Université)

Location: Aix-en-Provence, France

No. of positions: 1

Duration: 2 years (see below)

Application deadline: July 20, 2016

The Laboratoire Parole et Langage (CNRS, Aix-Marseille Université, Aix-en-Provence, FR) invites applications for one postdoctoral fellowship supported by a grant from the French National Research Agency (ANR). The funded project aims at exploring the role of prosody in speech planning and how planning strategies are affected both by the pragmatic situation and by individual cognitive-emotional abilities. Cognitive-emotional constraints are investigated by comparing healthy adults and clinically impaired individuals (patients affected by multiple sclerosis). The project involves an interdisciplinary team from across the Laboratoire Parole et Langage (which includes experts in linguistics, speech analysis, psychology) including collaboration with the 'Pôle de Neurosciences Cliniques' of the Hospital la Timone (Marseille) and the Neurology Department of the Centre Hospitalier du Pays d'Aix (Aix-en-Provence).

Applicants must have a Ph.D. in a relevant field of study (e.g., Cognitive Science, Linguistics, Psychology), and a strong background in experimental design and statistics. Special attention will be given to candidates with expertise in cognitive aspects of social coordination and in emotion in spoken language. Experience with psychophysiological methodologies and conducting studies with neurological patients is also highly desirable.

The postdoctoral appointment is for a duration of up to two years depending on the starting date, and subject to review after the first year. Applicants should hold a PhD or expect to complete a PhD by the starting date. The starting date is flexible, though the position should be filled by the beginning of the fall of 2016.

Applications should include (i) a cover letter that relates the applicants’ experience and interests to the project, (ii) a comprehensive CV, (iii) the names and contact information of two references, and (iv) a PDF copy of all publications or a list of links where these can be accessed.

Applications in French or English may be sent by email to Elisa Sneed German at elisa.sneed@lpl-aix.fr.

For further inquiries regarding the position or the project, please contact Caterina Petrone (Principal Investigator) at caterina.petrone@lpl-aix.fr.

Back

Top

=============================================
PhD Positions in Signal Processing for Speech Communication
=============================================

The Signal Processing and Speech Communication Lab (SPSC) of Graz University of
Technology is seeking to fill one PhD position in the area of signal processing for
speech communication.
The Ph.D. research will be embedded in the FWF project P 28070-N33 entiteld 'Phase-Aware
Signal Processing for Speech Communication' led by Dr. Pejman Mowlaee.

Required skills:
* M.Sc. degree in relevant field (Audio engineering, EE, CS, Telematics)
* Strong background in statistical signal processing and speech processing
* Independent and self-motivated working
* Excellent communication skills, fluency in English

The SPSC is a research and education center in focused on different aspects of signal
processing covering applications in speech/audio communication, telecommunications and
wireless communication. For more details on the ongoing research and publications carried
out in the project you are referred to PhaseLab webpage:
https://www2.spsc.tugraz.at/people/pmowlaee/PhaseLab.html.

Please send your application (CV, motivation letter, list of grades, and names of
references) to pejman.mowlaee@tugraz.at. Application deadline is August 1st, 2016.

Back

Top

Nous offrons un contrat doctoral de 3 ans portant sur la « Caractérisation et la génération de l'expressivité en fonction des styles de parole pour la construction de livres audio ». Des informations sur ce travail sont disponibles à : https://www-expression.irisa.fr/files/2016/06/phd_caracterisation_generation_livre_audio.pdf. Cette thèse sera effectuée en co-direction entre l?IRISA et le LLF (plus précisément LABEX EFL). 

Back

Top

http://www.dgdr.cnrs.fr/drhita/concoursita/consulter/resultats/consulter.htm

http://www.dgdr.cnrs.fr/drhita/concoursita/consulter/resultats/consulter.htm

Back

Top

Proposition de travaux de these, IRISA, equipe Expression

Titre : Optimisation de scripts d'enregistrement pour la lecture expressive de livres audio

Mots-clefs : synthese de la parole expressive ; optimisation et apprentissage.

Contexte : Le projet vise a etudier la realisation automatique de livres audio a l'aide d'une

voix de synthese. La duree d'ecoute de l'integralite d'une uvre necessite une voix de haute

qualite a l'expressivite adaptee.

Un systeme de synthese vocale a partir du texte (TTS) produit un signal de parole

correspondant a une vocalisation d'un texte donne. Ces dernieres annees, la TTS a fait de

nombreux progres en termes de qualite acoustique et d'intelligibilite, mais la production

d'une voix expressive de tres bonne qualite reste un verrou scientique (voir [1] et ses

references). Cette qualite vocale depend fortement du systeme TTS (parametrique, ou par

selection et concatenation d'unites sonores) et du corpus de parole utilise.

Frequemment, la creation d'un tel corpus de parole necessite l'enregistrement de la lecture

d'un script specique avec des types d'expressivite donnes. Ce processus d'enregistrement

etant complexe et co^uteux, de nombreux travaux portent sur la creation d'un script

permettant de couvrir un maximun d'evenements souhaites tout en minimisant sa duree

(voir [2, 3, 4] et leurs references).

Proposition : La problematique etudiee dans ce projet de these est la creation de livres

audio sous une forme hybride : il s'agit d'enregistrer une partie minimale des livres vises

pour produire une voix de synthese la mieux adaptee au texte restant a vocaliser. Plus

generalement, le sujet porte sur l'etude des methodes de construction et d'enrichissement

automatique de scripts d'enregistrement an de produire une voix de synthese de haute

qualite pour des textes pre-denis d'expressivite variee. Cette approche se formalise en un

probleme d'optimisation d'un compromis entre qualite des messages acoustiques naux et

quantite de texte a enregistrer.

Un premier axe de travail concerne la problematique de l'evaluation subjective et objective.

Dans le cadre general de la synthese de la parole, l'evaluation de la qualite des signaux

produits est un probleme qui fait l'objet de nombreuses etudes (voir par exemple [5, 6, 7])

mais qui reste dicile. En quoi le fait de conna^tre a l'avance le texte a vocaliser ou de

disposer de signaux de parole naturelle realises dans le m^eme contexte permet de simplier

ce probleme ? D'un autre cote, le livre audio produit sera un melange de signaux naturels

et de signaux de synthese. Il sera donc necessaire d'etudier et de proposer des approches

speciques pour evaluer de tels objets et, en particulier, depasser l'evaluation subjective a

l'echelle de la phrase.

Un deuxieme axe de travail porte sur la construction automatique du script d'enregistrement

et la denition d'un compromis entre la qualite des signaux et la taille de l'enregistrement

associe. Plusieurs verrous sont deja identies. Comment les descripteurs textuels

inuencent-ils la qualite nale ? En particulier, quelles methodes d'apprentissage, guidees

par des mesures objectives de qualite, conduisent aux jeux de descripteurs optimaux ?

Un dernier axe de travail porte sur l'etude de la prise en compte des alterations entre le

resultat theorique attendu lie au script d'enregistrement et le signal acoustique reel issu de

la phase d'enregistrement. Comment detecter ces variations et adapter dynamiquement le

script an de conserver la qualite acoustique nale initialement attendue ?

Environnement de travail : le projet sera realise au sein de l'equipe Expression de l'IRISA,

dans sa composante lannionnaise specialisee sur les problematiques de synthese de la parole

et de traitement automatique des langues. Il sera encadre conjointement par Damien Lolive

1

et Jonathan Chevelu (IRISA-ENSSAT Lannion, Universite de Rennes1) et beneciera d'un

nancement sur trois ans (nancement des conseils departemental et regional). L'equipe

dispose d'un moteur de synthese de la parole par corpus, d'un moteur statistique (HTS),

d'un studio d'enregistrement, d'une plate-forme de tests d'ecoute [8] et d'une collection de

livres audio annotes [9] qu'elle enrichit dans le cadre d'un projet ANR.

Prol du candidat : Le candidat sera dipl^ome d'un master informatique ou de toute autre

formation equivalente. Compte-tenu du sujet, des competences avancees en algorithmique

et programmation seront requises. Le candidat disposera de la motivation et des facultes

necessaires pour aborder les domaines de recherche de la synthese de la parole, de l'apprentissage

articiel et du traitement automatique des langues.

Contacts :

Damien LOLIVE (damien.lolive@irisa.fr) et Jonathan CHEVELU (jonathan.chevelu@irisa.fr)

Bibliographie

[1] D. Govind, S. R. Mahadeva Prasanna, Expressive speech synthesis : a review, Int. J. of

Speech Tech., p. 1-24, 2013.

[2] H. Francois, Synthese de la parole par concatenation d'unites acoustiques : construction

et exploitation d'une base de parole continue, these de l'Univ. de Rennes 1, 2002

[3] D. Cadic, Optimisation du procede de creation de voix en synthese par selection, these

de l'Univ. de Paris 11, 2011

[4] N. Barbot, O. Boe
ard, J. Chevelu, A. Delhay, Large linguistic corpus reduction with

SCP algorithms, Computational Linguistics 41(3) : 355-383, 2015

[5] N. Campbell, Evaluation of speech synthesis : from reading machines to talking machines,

Evaluation of Text and Speech Synthesis, (L. Dybjoer at al. Eds.) , Chapitre 2,

2007

[6] J. Chevelu, D. Lolive, S. Le Maguer, D. Guennec, How to compare TTS systems : a

new subjective evaluation methodology focused on di
erences, Interspeech, 2015

[7] C.-T. Do, M. Evrard, A. Leman, C. d'Alessandro, A. Rilliard, J.-L. Crebouw, Objective

evaluation of HMM-based Speech synthesis system using Kullback-Liebler divergence,

Interspeech, 2015

[8] L. Blin, O. Boe
ard, V. Barreaud, WEB-based listening test system for speech synthesis

and speech conversion evaluation, LREC, 2008

[9] O. Boe
ard, L. Charonnat, S. Le Maguer, D. Lolive, Towards fully automatic annotation

of audio books for TTS, LREC, 2012

2

Back

Top

Ingénieur(e) d’études – CDD 6 mois

Contexte et environnement

Dans le Projet MANES (Médiation et Accessibilité Numérique pour les Etudiants Sourds) financé par

l’UNR-RA), l’université Grenoble-Alpes recrute un ingénieur d’études pour un CDD de 6 mois (ou 2 CDD

de 3 mois).

L’objectif général du projet est de développer un dispositif de sous-titrage en temps-réel pour rendre

accessible le discours oral de l’enseignant aux étudiants sourds, de façon à favoriser l’appropriation

individuelle des savoirs, par le biais de la prise de notes. La réalisation technologique et les capacités de

traitement de l’écrit par les publics sourds seront les deux axes de ce projet.

L’ingénieur(e) d’études recruté(e) participera à l’étape « expérimentation du dispositif » visant à tester

l’apport d’un dispositif de ce type pour favoriser la prise de notes des étudiants sourds.

Missions principales

L’ingénieur(e) d’études recruté(e) sera chargé(e), en lien avec les chercheurs participant au projet, de

l’élaboration, de la passation et de l’analyse des expérimentations auprès des étudiants sourds.

o Montage d’une pré-expérimentation vidéo en ligne (plateforme moodle)

o Quizz vidéo LSF en ligne à réaliser : élaboration des questions, élaboration des vidéos

questions en LSF, élaboration d’un module de questionnaire en ligne sur moodle

o Annotation des prises de notes (français écrit) et des restitutions (LSF)

o adaptation du système de sous-titrage au public sourd

Compétences

- Expérience de recherche souhaitée dans le domaine de la surdité et/ou dans le domaine des TICE.

- Compétences en LSF et en français écrit

- Compétences linguistiques pour l’analyse du français écrit et de la LSF

- Compétences multimédia : outil de transcription (ELAN), de sous-titrage, de montage vidéo et

création de modules en ligne (MOODLE ou autre plateforme pédagogique).

- Une expérience en programmation web serait un plus.

Profil du candidat

- Issu(e) d’un master dans le domaine des Sciences du Langage ; Industrie de la Langue, Informatique,

Sciences de l’éducation, psychologie

Contexte de travail

L’ingénieur d’études travaillera au sein du laboratoire LIDILEM (Laboratoire de LInguistique et de

Didactique des Langues Etrangères et Maternelles) ou LIG (Laboratoire d’Informatique de Grenoble).

Modalités de candidatures

Les candidatures (CV détaillé et lettre de motivation) sont à adresser au porteur du projet MANES :

Isabelle Estève (Isabelle.esteve@univ-grenoble-alpes.fr ) ; dès à présent et jusqu’au vendredi 24 juin.

Back

Top

up to 4 PhD students in conversational systems and social robots

Back

Top

href='http://ipvs.informatik.uni-stuttgart.de/mlr/research-group-leader/'>
Research Group Leader in Intelligent Human-Robot Cooperation
< /a>
at University Stuttgart, Germany. Email Marc Toussaint
< marc.toussaint@informatik.uni-stuttgart.de>

Back

Top

Title: 1 PhD position in bio-inspired ASR, 1 PhD position in brain-signal based ASR

The Center for Translational Neurophysiology of Speech and Communication (CTNSC) at the Italian Institute of Technology is seeking for motivated PhD students who will work on bio-inspired automatic speech recognition (ASR) and brain-signal based ASR.

The successful candidates will have a Master degree in Computer Science, Engineering (or equivalent) and a background in machine learning and/or signal processing.

The candidates should apply by the 5^th of August, 2016.

PhD courses will start on the 1^st of November, 2016.

CTNSC Info: https://www.iit.it/centers/ctnsc-unife

PhD page [ITA]:

http://www.unife.it/studenti/dottorato/concorsi/ordinario

PhD course summary [ITA]:

http://www.unife.it/studenti/dottorato/modulistica/concorsi/32-ciclo/neuroscienze

How to participate to the selection [ITA/ENG]:

http://www.unife.it/studenti/dottorato/modulistica/Guidaconcorso.pdf

Back

Top

Job offer: permanent research position in machine learning for natural language processing

Mission: Conceive algorithms and develop softwares for natural language processing, with applications to automatic extraction of structured knowledge from documents

Position environment: In Orange Labs,  the « CONTENT » department prepares the future of content access services (music, video, press, radio, education, entertainment). The job is opened in a multidisciplinary team, located in Lannion, France, in which are conducted collaborative studies with academics (collaborative projects, PhD supervision, scientific publications) and research and development projects with strong interactions with operational units of Orange (marketing, customer experience, design?)  

Candidate profile: PhD or significant research experience in the field of machine learning, with an noticeable interest on natural language processing

In order to visualize the full offer and to apply on line : https://orange.jobs/jobs/offer.do?joid=54926&lang=FR

For more information, please contact:

Gilles Prigent (gilles.prigent@orange.com)

Delphine Charlet (delphine.charlet@orange.com)

Géraldine Damnati (geraldine.damnati@orange.com)

Back

Top

ASR Research Scientist (Chinese) Oben Pasadena, California

  
Founded in 2014, ObEN is an artificial iintelligence company based in one of the world’s most successful incubators: Idealab in Pasadena, CA. We have developed a sophisticated speech technology that allows any Internet of Things to speak in any voice and in any language. 

 
JOB DESCRIPTION  As a Speech Recognition Research Scientist at ObEN, you’ll be working on several proprietary and stealthy projects in the audio space. Your work will encompass the whole scope of application development, including speech research, voice interface design, application prototyping, and development of our proprietary speech recognition engine. Currently, we are mainly interested in candidates with experience in ASR for Chinese language. Experience in English and/or Spanish is a big plus.  
You must have:

■ PhD in Computer Science, Electrical Engineering or Mathematics with specialization in speech recognition, natural language processing or machine learning.

■ Solid experience building ASR systems (Chinese) and/or a publication record in the area.

■ Solid experience with Kaldi.

■ Strong machine learning background and familiar with standard statistical modeling techniques applied to speech.

■ Familiarity with linguistic phonetics.

■ Proficiency in programming languages such as C/C++, Python, Java or Perl.

■ Knowledge of basic digital signal processing techniques for audio.

■ Enjoys a highly collaborative environment. 

To apply please send the following: to careers@oben.me

● Detailed resume and/or Linkedin profile.

● Links to prominent scientific/professional contributions.

Contact: Fernando Villavicencio <fernando@oben.com>

Back

Top

In the context of a Swiss NSF grant, we seek a PhD student to work on multilingual and
affective speech synthesis.

  http://www.idiap.ch/education-and-jobs/job-10193

The research will begin with the state of the art in speech synthesis, probably based
around deep neural networks (DNNs). Such networks have been shown recently to produce
synthetic speech with better overall quality, especially with respect prosody, than other
techniques. The student will then focus on modelling of speech with a view to extraction,
modeling and adaptation of emotion (affect) in synthetic speech. To demonstrate the
language independence of the models, it will be necessary to work in two or three
languages.

Several research threads are possible; one thread will involve creating a neural model of
prosodic features. It will build upon recent work here at Idiap where we have created a
general physiologically plausible model for prosody.

Another research thread will involve analysis and modification of the phonetic quality of
the synthetic speech. This is in keeping with observations that formant positions can
change with affective indicators such as valence and arousal.

Depending on progress, there is also the possibility to address cross-lingual modelling
of affect. Here, the goal would be to detect emotion in one language and reproduce it in
another language.

Owing to availability of data, the work will probably proceed using German data. However,
our locality dictates that evaluation will be easier in French. Amongst the Swiss
languages, Italian is also appealing.

The ideal Ph.D student should have a master (or equivalent) degree in engineering,
computer science, or applied mathematics. S/he should have a good background in
mathematics, statistics, and programming (C/C++, Python, scripting langages). Given the
multlingual nature of the project, the position would suit someone with some knowledge of
one or more of the languages described above. In order to balance the group, we
especially encourage female applicants. However, all applications will be judged on merit.

Applications will be considered as they are received; the position will remain open until
filled.

About Idiap:

Idiap is an independent, non-profit research institute recognized and supported by the
Swiss Government, and affiliated with the Ecole Polytechnique Fédérale de Lausanne
(EPFL). It is located in the town of Martigny in Valais, a scenic region in the south of
Switzerland, surrounded by the highest mountains of Europe, and offering exciting
recreational activities, including hiking, climbing and skiing, as well as varied
cultural activities. It is within close proximity to Geneva and Lausanne. Although Idiap
is located in the French part of Switzerland, English is the working language. Free
French lessons are provided.

Idiap offers competitive salaries and conditions at all levels in a young, dynamic, and
multicultural environment. Idiap is an equal opportunity employer and is actively
involved in the 'Advancement of Women in Science' European initiative. The Institute
seeks to maintain a principle of open competition (on the basis of merit) to appoint the
best candidate, provides equal opportunity for all candidates, and equally encourage both
genders to apply.

--
Phil Garner
http://www.idiap.ch/~pgarner

Back

Top

Speech Analytics Engineer

Job Description

SRI’s Speech Technology and Research (STAR) Lab seeks a self-motivated and team-oriented speech engineer to join a team working on modeling of speaker state (for example, emotion and health). The position offers both research and commercial opportunities, and a potential for growth and client interaction. The work will include advancing technology capabilities, understanding client data and conducting experiments involving model development and testing, under current or new projects in this area. This is an opportunity to have a significant impact in an emerging research area.

STAR Lab engages in leading-edge research in speech recognition, speaker characterization, speaker and language identification, machine translation, natural language processing and other areas of speech/language technology, offering opportunities from basic research to prototyping, productization, and widespread deployment. The multidisciplinary research team consists of excellent speech researchers, linguists and software engineers. Characteristics of STAR staff are enthusiasm, self-motivation, initiative, passion for learning, taking ownership, thriving in a flat organizational hierarchy, and a desire to control their own career paths.

SRI International is a premier nonprofit research organization based in Menlo Park, California with a mission to create world-changing solutions making people safer, healthier, and more productive.

Located in the heart of Silicon Valley, the world’s center for innovation and technology, SRI’s focus spans everything from basic research to delivered systems, government and commercial work and spinning off successful companies like SIRI. Flexible cross-laboratory teams often form dynamically to solve challenging problems spanning multiple disciplines.

Requirements

-PhD (preferred) or Master’s degree in electrical engineering or computer science, with experience in advanced speech signal processing and in state of the art machine learning techniques for speech processing.

-Additional background in natural language processing is useful.

-Work experience outside the PhD institution, including internships, is preferred.

-Candidates should have an interest in the topic area, excellent programming skills, be a quick learner, be proactive and efficient, and work well both individually and in teams.

Apply online at www.sri.com/careers

Job ID: 1239BR

SRI is an Equal Employment Opportunity/Affirmative Action Employer

Back

Top

Announcing 14 PhD positions on a new Marie Curie European Training Network, ENRICH (Enriched Communication Across the Lifespan)

Speech is a hugely efficient means of communication: a reduced capacity in listening or speaking creates a significant barrier to social inclusion at all points through the lifespan, in education, work and at home. Hearing aids and speech synthesis can help address this reduced capacity but their use imposes greater listener effort. The fundamental objective of ENRICH is to modify or augment speech with additional information to make it easier to process. Enrichment aims to reduce the listening burden by minimising cognitive load, while maintaining or improving intelligibility. ENRICH will investigate the relationship between cognitive effort and different forms of natural and synthetic speech.

ENRICH is funded for 4 years from October 2016. We are now looking to recruit 14 generously-funded early stage researchers to undertake projects leading to a PhD within the network. Openings exist for candidates with a background in engineering, psychology, speech science or linguistics.

ENRICH partners are: University of Edinburgh, Radboud University, Tobii Technology, University College London, University of Crete, Hoerzentrum Oldenburg, University Medical Center Groningen and the University of the Basque Country (coordinator).

For more details of the network and early stage researcher positions, see http://www.enrich-etn.eu. The closing date for applications is 30 September 2016.

Back

Top

Le service audio de la Police Technique et Scientifique (Ecully, près de Lyon) recherche des vacataires pour effectuer un travail de segmentation et/ou d'enregistrements de locuteurs. 
Le profil suivant est recherché:

ou de téléphoner au 0472868638

Back

Top