Modeling Virtual Coaches

The European Horizon-2020 project Council of Coaches aims to develop a tool in which virtual embodied coaches form a team of experts that can discuss amongst themselves and with the user how the user could improve his healthy lifestyle behaviors. The project combines state of the art 3D Virtual Avatars with language and reasoning technology and applies this to the area of lifestyle and behavior change coaching.

The goal of research is to design, implement and evaluate user interaction with the team of virtual agents, the council of coaches (Home Interface) and one-on-one interaction with a single coach (Mobile Interface). In particular it focuses on

1. Modelling coaches with their own specific agenda and behaviour style aiming at creating unique social relationships with users.
2. Adapting virtual coaches? behavior to ensure that user maintains his involvement in the relationship with the coaches.

Specific attention will be given to the development of virtual coaches with a wide variety of behavioural characteristics and personality traits. Several studies have shown how low level features such as behaviour expressivity and high level characteristic (e.g. personality traits ) affect user?s involvement in their interaction with the virtual characters and impact his performances (McRorie et al., 2011) (Paiva et al., 2017)).

The work in this research will make use of and build upon the GRETA/VIB platform developed at UPMC (Pecune et al., 2014) for multimodal behaviour generation and for visualizing virtual coaches.

To apply, send a CV and names of reference to    catherine.pelachaud@upmc.fr

H2020 Council of Coaches

Context:

The GRETA/VIB platform is developed at CNRS-ISIR (Pécune et al, 2014). It simulates virtual characters able to communicate with humans in real-time. It is endowed with socio-emotional capabilities. The control of the character is done through two specific languages at the communicative intention level and at the multimodal behavior one. The platform includes several tools to create multimodal behaviors.

The GRETA/VIB platform is used within the European Horizon-2020 project Council of Coaches which aims to develop a tool in which virtual embodied coaches form a team of experts that can discuss amongst themselves and with the user how the user could improve his healthy lifestyle behaviors. The project combines state of the art 3D Virtual characters with language and reasoning technology and applies this to the area of lifestyle and behavior change coaching.

Job Description:

We are looking for an engineer knowledgeable in 3D virtual environment. S/he will participate to Council of Coaches project. Her/his role will be:

-          Port the Greta/VIB platform onto Android mobile and VR headset.

-          Participate to development of computational model of agent?s multimodal behaviors

-          Integrate software modules within the VIB/Greta platform and/or within a bigger software system developed within the H2020 project Coach Council.

-          Participate to developers meetings (online and on site) of the H2020 project Coach Council

Profile: Engineer in computer science, Master, PhD

skills:

• Programming languages and standards: Java, C#, C++, XML, Ogre 3D

Mastered skills:

• The ideal candidate will have solid software development, testing and integration skills
• English (written and oral)

Project Length: 1 year position renewable

Place: ISIR - UPMC

Stipend: depends on applicant qualification

Contact: Catherine Pelachaud, CNRS-ISIR; catherine.pelachaud@upmc.fr

To apply, send a CV, names of reference, master grades (for Master applicants) to    catherine.pelachaud@upmc.fr

Top

The Institute of Digital Sciences, Management and Cognition at the University of Lorraine
(Nancy, France) opens a new Masters program in Natural Language Processing and Data
Science - Computer Science, Speech, Language, and Knowledge Representation

http://institut-sciences-digitales.fr/idmc-master-degree-in-natural-language-processing/

Do you want to become an expert in Deep Neural Networks, Logic, Speech Processing,
Information Retrieval, Knowledge Representation, ? all of this for Natural Language?
Here's your chance!

Natural Language Processing (NLP) is at the crossroads of linguistics, computer science,
and artificial intelligence. We propose a modern curriculum which combines different
approaches covering both theoretical and applied perspectives.

The program includes a hands-on project in each semester. It ends with a 6-month paid
internship in a company or a research lab.

See the course description below.


----------------------------------------------------------------------------------------
You can apply for the 1st year (M1) or for directly entering the 2nd year (M2).
----------------------------------------------------------------------------------------

Language
-------------
All courses are taught in English (except the 'French for non-native speakers' course).


Fees
-------
The University of Lorraine is publicly funded and therefore offers tuition-free education
for all students including students from both inside and outside EU/EEA/EFTA countries.
The only student expenditures are a nominal semester fee of about 600 euros which
includes health insurance. Nancy?s high quality of life goes hand-in-hand with a low cost
of living.


Why Nancy?
---------------
Nancy is a center of excellence for NLP!

1/Two important research centers:
- LORIA, the computer science research laboratory (http://www.loria.fr/en/)
- ATILF the linguistics research laboratory (http://www.atilf.fr/)

Both research units participate in many national and international research projects.
Their members are highly visible in the international community and regularly publish in
the most prestigious conferences and journals in the field.

2/ The University of Lorraine participates in two Erasmus Mundus Masters funded by the
European Union :
- Erasmus Mundus Master Language and Communication Technology (http://lct-master.org/)
- Erasmus Mundus Master EMLex (https://www.emlex.phil.fau.eu/)


Why NLP?
--------------
Natural Language Processing is a rapidly expanding field.

With the proliferation of digital data, there is a massive need for well-trained
engineers and researchers able to exploit this data for commercial (e-commerce,
recommendation systems, translation, etc) and socio-political (e-learning, opinion
mining, behavioral prediction, etc.) purposes.

Start-ups, small firms with strong R&D profiles and large companies (e.g., in France,
Airbus, Thales, EDF, Orange) are looking for scientists and engineers who can contribute
to the development of systems capable of managing and exploiting textual data.

The GAFAM (Google, Amazon, Facebook, Apple, Microsoft) are hiring intensively in both the
NLP and the Speech processing domains.

The need for PhD candidates is also strong as shown by the many PhD scholarships offered
by both academic and industrial players.


Contacts
------------
Head of the MSc program:
        Maxime Amblard <maxime.amblard@univ-lorraine.fr>
Secretary:
        Karine Weisse <karine.weisse@univ-lorraine.fr>
Location of the courses :
        École des Mines (Artem Campus, Tram stop 'Blandan')
        http://www.alliance-artem.fr/


Application
---------------
The application file is available here:
https://ecandidat.univ-lorraine.fr

More information:
http://institut-sciences-digitales.fr/application-process-nlp-master-degree/#application


Academic Year 2018-2019
-----------------------------------
From September 3rd to 13th, 2018: Refresher courses in mathematics (probability theory,
statistics) and computer science (algorithms, programming)

M1:
        1st semester from September 17th to January 11th
        2nd semester from January 14th to June 6th
M2:
        1st semester from September 17th to February 15th
        Internship from February to August


In brief
----------
? A program with strong national and international visibility
? Taught entirely in English
? A high quality research environment with two research centers covering a wide range of
disciplines (linguistics, computer science, speech and language processing) and
applications (Human-Machine Interfaces, Speech Processing, Knowledge Retrieval,
Statistical Machine Translation, etc.)
? Studies in a vibrant, human-sized city
? Studies at a low cost (Tuition fees at around 600 euros per year incl. health insurance)


Course Description
---------------------------
M1 1st Semester
---------------------------
PROBABILITIES, STATISTICS AND ALGORITHMS FOR AI
        ? Elementary mathematical tools, statistics, algorithms to define and solve
artificial intelligence problems
        ? Case studies covering both theoretical and practical aspects

CORPUS DESIGN AND ACQUISTION
        ? Written corpora
        ? Speech corpora

SOFTWARE ENGINEERING
        ? Collection, analysis and formalization of customers' needs
        ? Software design and development
        ? Functional analysis and specifications
        ? Project management

LINGUISTICS FOR NLP (1)
        ? Methods for Natural Language Processing
        ? Phonology
        ? Morphology

PROJECT AND FOREIGN LANGUAGE
        ? Hands-on project
        ? Foreign language (French)

---------------------------
M1 2nd Semester
---------------------------
MACHINE LEARNING AND SEMANTIC WEB
        ? Machine learning theory
        ? Data mining
        ? Knowledge retrieval

FORMAL TOOLS
        ? Logic
        ? Formal languages
        ? Calculability and complexity

DATA PROCESSING
        ? Storage and retrieval
        ? Data analysis

LINGUISTICS FOR NLP (2)
        ? Lexicology: lexical units and phraseology
        ? Syntax
        ? Semantics

PROJECT AND FOREIGN LANGUAGE
        ? Hands-on project
        ? Scientific communication
        ? Foreign language (French)

---------------------------
M2 1st Semester
---------------------------
DEEP LEARNING AND DATA MINING
        ? Neural networks, Deep neural networks
        ? Data mining (structured data and text)
        ? Collaborative filtering

TEXT AND SPEECH PROCESSING
        ? Processing textual data
        ? Speech processing
        ? Terminology and ontology

NATURAL LANGUAGE PROCESSING AND DISCOURSE
        ? Application to text
        ? Computational semantics
        ? Discourse and Dialog modeling

LEXICONS AND GRAMMARS FOR NLP
        ? Diachronic and synchronic lexicology
        ? Lexical resources
        ? Syntactic framework

PROJECTS AND FOREIGN LANGUAGE
        ? Software project
        ? Law and ethics
        ? Research methods
        ? Professional integration
        ? Foreign language (French)

---------------------------
M2 2nd Semester
---------------------------
Paid internship (company or research lab)

Top

M.Sc. Program in Natural Language Processing (NLP) and Data Science
Université de Lorraine, Nancy (France)


The Institute of Digital Sciences, Management and Cognition
is opening a Masters Program in NLP - ??
Computer Science, Speech, Language and Knowledge Representation

**************

http://institut-sciences-digitales.fr/idmc-master-degree-in-natural-language-processing/
**************

So you want to be a specialist in Neural Networks, Logic, Speech
Processing, Information Retrieval, Knowledge Representation... all
for Natural Language? Well, now's your chance!

Natural Language Processing (NLP) lies at the crossroads of
linguistics, computer science and artificial intelligence. This
Masters Program offers a modern curriculum which combines the
different approaches, and covers both theoretical and applied
perspectives.


In each semester, the program includes a hands-on project.
It ends with a 6-month paid internship in a company or a research
lab. You can find the course description below.


----------------------------------------------------------------------------------------
You can apply to directly enter at either the first year (M1) or second year (M2) level.
----------------------------------------------------------------------------------------

Language
-------------
All courses are taught in English
(except the 'French for non-native speakers' class).


Fees
-------
The University of Lorraine is publicly funded and thus offers
tuition-free education for many students including students
from the EU (243 ? cost per year). The tuition fee for students
from other nations arriving in France at the beginning of
the 2019 academic year will be 3.770 ? per year.

Find more information here:
https://www.campusfrance.org/en/tuition-fees-France

Nancy?s high quality of life goes hand-in-hand with a low cost
of living.


Why Nancy?
---------------
Nancy is a center of excellence for NLP. It can boast of:

1/Two important research units:
- LORIA, the computer science research laboratory http://www.loria.fr/en/
- ATILF the linguistic research laboratory http://www.atilf.fr/

Both research units participate in many national and
international research projects. Their members are highly
visible in the international community and regularly publish
in the most prestigious conferences and journals of the field.

2/ Two Erasmus Mundus Masters funded by the European Union :
- Erasmus Mundus Master Language and Communication
Technology (LCT) http://lct-master.org/
- Erasmus Mundus Master EMLex.
        https://www.emlex.phil.fau.eu/


Why NLP?
--------------
Natural Language Processing is a rapidly expanding field.

With the proliferation of digital data, there is a massive need
for well-trained engineers and researchers able to exploit this
data for commercial (for example: e-commerce, recommender
systems and translation) and socio-political (for example:
e-learning, opinion mining, and behavioural predictions) purposes.

Start-ups, small firms with strong R&D profiles and large companies
(in France this includes Airbus, Thales, EDF, Orange, etc.) are
looking for scientists and engineers who can contribute to the
development of systems capable of managing and exploiting textual
data.

The GAFAM (Google, Amazon, Facebook, Apple, Microsoft) also recruit
intensively in both the NLP and the Speech processing domain.

There is also a strong need for Ph.D. candidates in the field, as
demonstrated by the many Ph.D. proposals in both the public and
the industrial domain.



Contacts
------------
Head of the Master program: Maxime Amblard
        maxime.amblard@univ-lorraine.fr
Secretary: Karine Weisse
        karine.weisse@univ-lorraine.fr

Location of the courses:
        École des Mines (Artem Campus, Tram stop 'Blandan')
        http://www.alliance-artem.fr/


Application
---------------
Non-European students
---------------
Students from most non-UE countries shall apply on
https://pastel.diplomatie.gouv.fr/etudesenfrance/dyn/public/authentification/login.html?_csrf=a530f0a6-e1d3-4052-ba91-f2cb8029fcd8&codeLangue=EN
(search Traitement Automatique des Langues Nancy)

Students from other non-EU countries can directly apply with the following file:
http://institut-sciences-digitales.fr/applicationfilem-sc_-nlp19-20/


European students
---------------
The application file for EU students is available here:
https://ecandidat.univ-lorraine.fr

Create an account then follow:
        - Offre de formation
        - Institut des sciences du digital, management et cognition
        - Master
        - M1-Traitement Automatique des Langues (TAL) (NANCY) (FI)
           or
        - M2-Traitement Automatique des Langues (TAL) PT Informatique, langue, parole et
connaissances (NANCY) (FI)

More information:
http://institut-sciences-digitales.fr/application-process-nlp-master-degree/#application


Academic Year 2019-2020
-----------------------------------
From 3rd to 13th September 2019: Refresher courses in
mathematics (probability theory and statistics) and computer
science (algorithms and programming).

M1:
        1st semester from September 2nd to January 10th
        2nd semester from January 13th to June 5th
M2:
        1st semester from September 2nd to March 27th
        Internship from February to August


In brief
----------
? A program with strong national and international visibility
? Taught entirely in English
? A high quality research environment with two research centers
covering a wide range of disciplines (linguistics, computer
science, speech and language processing) and applications
(human-machine interfaces, speech processing, knowledge
extraction, statistical machine translation, etc.)
? Studies in a vibrant, human-size city


Course Description
---------------------------
M1 Semester 1
---------------------------
PROBABILITIES, STATISTICS AND ALGORITHMS FOR AI
        ? Elementary mathematical tools, statistics, algorithms
        to define and solve artificial intelligence problems
        ? Case study approach towards mastering both theoretical
        and practical aspects of a topic

DESIGN AND ACQUISITION OF CORPORA
        ? Written corpora
        ? Speech corpora

SOFTWARE ENGINEERING
        ? Collecting, analysing and formalizing customers' needs
        ? Software design and development
        ? Functional analysis and specifications
        ? Project management

LINGUISTICS FOR NLP (1)
        ? Methods for Natural Language Processing
        ? Phonology
        ? Morphology

PROJECT AND FOREIGN LANGUAGE
        ? Hands-on project
        ? Foreign language (French)

---------------------------
M1 Semester 2
---------------------------
MACHINE LEARNING AND SEMANTIC WEB
        ? Machine learning theory
        ? Data mining
        ? Knowledge extraction

FORMAL TOOLS
        ? Logic
        ? Formal languages
        ? Calculability and complexity

DATA PROCESSING
        ? Storage and retrieval
        ? Data analysis

LINGUISTICS FOR NLP (2)
        ? Lexicology : lexical units and phraseology
        ? Syntax
        ? Semantics

PROJECT AND FOREIGN LANGUAGE
        ? Hands-on project
        ? Scientific communication
        ? Foreign language (French)

---------------------------
M2 Semester 1
---------------------------
DEEP LEARNING AND DATA MINING
        ? Neural Networks, Deep Neural Networks
        ? Data mining (structured data and text)
        ? Collaborative filtering

TEXT AND SPEECH PROCESSING
        ? Processing Textual Data
        ? Speech processing
        ? Terminology and ontology

NATURAL LANGUAGE PROCESSING AND DISCOURSE
        ? Application to text
        ? Computational semantics
        ? Discourse and Dialog modeling

LEXICONS AND GRAMMARS FOR NLP
        ? Diachronic and synchronic lexicology
        ? Lexical resources
        ? Syntactic framework

PROJECTS AND FOREIGN LANGUAGE
        ? Software project
        ? Law and ethics
        ? Research methods
        ? Professional integration
        ? Foreign language (French)

---------------------------
M2 Semester 2
---------------------------
Paid internship (company or research lab)

Top

FIAT/IFTA Media Study Grants

The Media Studies Commission of the International Federation of Television Archives (FIAT/IFTA) is dedicated to fostering collaboration between research and archive communities and mediating the growth of scholarly expertise that adds value to audiovisual archives by means of innovative research. To this purpose, the Media Studies Commission has set up the Media Studies Grant as a way to promote and ensure the valorization of academic knowledge for archival practice. It is a programme that offers support for research carried out at FIAT/IFTA member archives or is of direct relevance to one or more of our member archives. Priority is given to projects that are relevant for the history of member archive institutions, or promise innovative insights into (digital) media historiography or archival practice in general.

2020 Call for Projects

In 2020, FIAT/IFTA’s Media Studies Commission is looking to commission research that adds value to and helps us understand the role of audiovisual archives in a shifting, converging media environment.

Digitization and digital tools enable novel ways of doing and telling media histories. With it, new possibilities for working with archival material and data in academic research open up. At the same time, new tools for discovery, annotation, visualization now span the possibilities of analyzing cultural heritage with a ‘long data’ and distant reading perspective We welcome studies addressing (but not necessarily limited to) the following areas:

1. Tools and (digital) methods: Producing new narratives in media history by means of digital tools and experimental methods of research and displaying different ways of doing (digital) media history as a way to redefine the value of historical archival content.

2. Storytelling: Highlighting new and interactive forms of storytelling that help revisit media history practices in the digital age.

3. Archives as (big) data: Reconceptualizing archives as (big) data to help us tap into unexplored narratives of media history. Reframing the historical and public value of audiovisual archives with mining and visualization tools.

4. Archival access in the digital age: Renewing strategies and practices for making archives accessible through different digital spaces, including social media and mobile apps. Addressing archival access strategies in an international context or working with datasets¹ from different archives.

5. Users and use practices: Revisiting approaches for researching users and user practices in relation to digital archives.

Awarded candidated are expected to deliver by the end of their grant period:

• A written report of their research at the quality standards of a scholarly article.

• An audiovisual essay, an interactive digital story or a creative demo (depending on the nature of the proposed project). Please make sure there are no copyright restrictions for the archival material you may want to re-use in this type of output.²

All output needs to mention the support of FIAT/IFTA. Successful candidates are required to present their research results at the FIAT/IFTA World Conference in Dublin, 26-29 October 2020.

Objectives

The Media Studies Grant makes available a maximum of €7000 for original and innovative projects that aim to:

• help us understand the role of audiovisual archives in a digital environment.

• unearth new narratives of media history.

• collect, disseminate and engage with historical sources that complement existing archive collections.

• conceptualize new ways of doing media history in the digital age.

• experiment with innovative methods of researching (digital) archives.

• valorize and disseminate knowledge of archive collections by means of interactive publications, multi-media presentations, digital applications and other creative demos.

• Produce open-source softwares aimed at a better understanding and exploitation of audiovisual archives.

• Release open-data corpora that would contribute to help further media studies.

Any questions? Please contact our commission members! For questions pertaining to access, archival collections and datasets made available for reasearch by different archival institutions, you may contact our following members:

Herbert Hayduck (ORF, Austria): archiv@orf.at
Claude Mussou (INA, France): cmussou@ina.fr
Silvia Proscia (RAI, Italy): silvia.proscia@rai.it
Masaya Maéda (NHK, Japan): maeda.m-is@nhk.or.jp
Daniela Floris (RAI, Italy): daniela.floris@rai.it
David Doukhan (INA, France): ddoukhan@ina.fr
Virginia Bazán Gil (RTVE, Spain): virginia.bazan@rtve.es
Lisa Kerrigan (BFI, UK): Lisa.Kerrigan@bfi.org.uk
Bas Agterberg (Netherlands Institute for Sound and Vision, NL): bagterberg@beeldengeluid.nl
Liam Wylie (RTÉ, Ireland): Liam.Wylie@rte.ie

Questions pertaining to academic research can be addressed to:

Dana Mustata, University of Groningen, D.Mustata@rug.nl
Andy O’Dwyer, University of Luxembourg, andy.odwyer@uni.lu

Requirements

• Candidates are required to send in their application in PDF format by 15 March 2020.

• Applications should be emailed to: Herbert Hayduck, archiv@orf.at

• Candidates are required to present their findings at the FIAT/IFTA World Conference and deliever all output as agreed upon in the funding agreemenent between FIAT/IFTA and the candidate.

• All results, publications and presentations derived from the study should mention the support of FIAT/IFTA Media Studies Commission and should be made available to FIAT/IFTA.

• FIAT/IFTA reserves the right to make accessible the output of funded studies on its own website.

• Proposed studies can be part of a bigger project (e.g. master thesis, PhD dissertation, book project, etc.) or can be stand-alone research initiatives that the candidate wishes to pursue.

Budget

Applicants can ask for a budget of €2500 up to a maximum of €7000 to support their travel and accommodation costs for the purpose of the proposed study, including travel to FIAT/IFTA-event (see below under “output”).

Researchers affiliated to a research institution who also have teaching responsibilities can use (part of) the allocated budget to buy themselves the research time needed for the proposed study. In this case, they should specify the research time that will be charged on the budget.

Senior researchers may use (parts of) the budget to hire interns or student assistants to assist with tasks on the project (e.g, corpus annotation). When third-party personnel is hired on the project, the senior researcher remains in charge of the project and is resonsibible for the final output.

Eligibility

We encourage master and PhD students as well as researchers affiliated to a university to apply for a Media Studies Grant with FIAT/IFTA Media Studies Commission.

Master and PhD students applying for a Study Grant need to send in together with their application a letter from their thesis supervisor showing support for the proposed study.

Awarded studies must be affiliated to a FIAT/IFTA member archive, either by exploring their collection or datasets for research purposes and/or carrying out research that is of direct relevance to a member archive institution.

Applicants should provide a support letter from FIAT/IFTA member‘s representative providing the collections which will be used in the study.

How to apply:

Applicants must send in an application for a Media Studies Grant. The applications should not exceed a maximum of 1000 words and need to include a:

Project description. This should outline the rationale of the project, the main research question(s), a description of the topic being researched, indication of the archival material or dataset to be studied and an explanation of the relevance of the proposed study for the research as well as archive communities. In case the project is part of a larger project, the candidate needs to specify how the proposed study contributes to the overall project.

Output. Candidates should specify the output resulting from the proposed study, including publications, presentations, software development, corpora annotation, and any other forms of knowledge utilization (e.g. virtual exhibitions, video essays, interactive storytelling applications etc.). At the end of their study, successful candidates are required to present their findings at a FIAT/IFTA public event to be agreed between the Media Studies Commission and the successful candidate and send in a written report outlining the research they’ve conducted and the main findings of their research. This report will be made available on the FIAT/IFTA website.

Workplan. Applicants should detail as much as possible all the research activities they plan to carry out as part of the proposed study and the time allocated for each of these activities. Please be aware that the Media Studies Grant only supports small-scale projects of 3-5 months.

Budget. Travel, accommodation and secondment costs can be covered by the Media Studies Grant. The budget can also be used for funding short-term internships, archive annotation campaigns, or paying for transport/accomodation fees required for the interactions with the FIAT archive member. For experimental projects, minimal technical costs may be eligible. In any technical costs are involved, we encourage the applicant to make contact with the Media Studies Commission before submitting their application, to ensure that the costs can be covered by the grant. Applicants should detail how the budget will be used. In case the proposed study is part of a bigger project, the applicant should specify any additional funding he/she may have received. Researchers asking for a secondment grant, should specify how their allocation of teaching and research time is divided and how much research time they wish to allocate to the proposed study and charge on the budget. Senior researchers hiring interns or student assistants to assist with the project, need to provide a statement with the number of hours and the budget allocated to the hired personnel.

Communication and dissemination activities. Applicants should present a communication and dissemination plan of how they intend to publicize and make available the findings of their study for the research and archive communities.

Applications should be sent in PDF format by 15 March 2020 to Herbert Hayduck at: archiv@orf.at.

All applications will be assessed internally by the members of the Media Studies Commission. Selected candidates are expected to present their study at the International FIAT/IFTA World Conference in Dublin.

More about the International Federation of Television Archives (FIAT/IFTA)

FIAT/IFTA provides a forum for exchange of knowledge and experience between its members, to promote the study of any topic relevant to the development and use of audiovisual archives and to establish international standards on key issues regarding all aspects of audiovisual media archive management. Within FIAT/IFTA, the Media Studies Commission promotes academic research that promotes knowledge, understanding and research of holdings of member audiovisual archives.

Top

ISCApad INTERSPEECH March 2020

March 10, 2020

The Fearless Steps Challenge (Phase 2: FS#2)

TIMELINE: Challenge Start Date (Data Release): January 25th 2020

INTERSPEECH-2020 Papers dealing with FEARLESS STEPS deadline: May 8, 2020

Registration Link:https://bit.ly/2qZ5tic

Challenge Tasks in Phase-2 (FS#2):

1. Speech Activity Detection (SAD)

2. Speaker Identification (SID)

3. Speaker Diarization:

3a. Track 1: Diarization using reference SAD

3b. Track 2: Diarization using system SAD

4. Automatic Speech Recognition (ASR):

4a. Track 1: ASR using reference Diarization

4b. Track 2: Continuous stream ASR

Website Link

https://fearless-steps.github.io/ChallengePhase2/

Background:

The Fearless Steps Initiative by UTDallas-CRSS led to the digitization, recovery, and diarization of 19,000 hours of original analog audio data, as well as the development of algorithms to extract meaningful information from this multichannel naturalistic data resource. As an initial step to motivate a stream-lined and collaborative effort from the speech and language community, UTDallas-CRSS is hosting a series of progressively complex tasks to promote advanced research on naturalistic “Big Data” corpora. This began with ISCA INTERSPEECH-2019: 'The FEARLESS STEPS Challenge: Massive Naturalistic Audio (FS-#1)'. This first edition of this challenge encouraged the development of core unsupervised/semi-supervised speech and language systems for single-channel data with low resource availability, serving as the “First Step” towards extracting high-level information from such massive unlabeled corpora.

As a natural progression following the successful Inaugural Challenge FS#1, the FEARLESS STEPS Challenge Phase-#2 focuses on development of single-channel supervised learning strategies. This FS#2 provides 80 hours of ground-truth data through Training and Development sets, with an additional 20 hours of blind-set Evaluation data. Based on feedback from the Fearless Steps participants, additional Tracks for streamlined speech recognition and speaker diarization have been included in the FS#2. The results for this Challenge will be presented at the ISCA INTERSPEECH-2020 Special Session. We encourage participants to explore any and all research tasks of interest with the Fearless Steps Corpus – with suggested Task Domains listed below. Research participants can however, also utilize the FS#2 corpus to explore additional problems dealing with naturalistic data, which we welcome as part of the special session.

Organizers

John H.L. Hansen (john.hansen@utdallas.edu)

Aditya Joglekar (aditya.joglekar@utdallas.edu)

Meena Chandra Shekar (meena.chandrashekar@utdallas.edu)

Abhijeet Sangwan (abhijeet.sangwan@utdallas.edu)

Top

ICASSP 2020 registration has re-opened! Sign up to view recorded on-demand content through June 8, 2020.

Top

Dear Sir or Madam,

We are contacting you in the context of a research project on the development of standard reading passages for speech and voice assessment. This study is part of Timothy Pommée’s PhD thesis at the Institut de Recherche en Informatique de Toulouse (Université Toulouse III Paul Sabatier, France). This project is carried out in collaboration with the Hospitals of Toulouse, the University Cancer Institute of Toulouse Oncopole and the University of Oslo. The primary aim of this thesis is to identify the needs for objective assessment of speech disorders (particularly speech intelligibility), in order to help develop new speech assessment tools that are tailored to the needs of clinicians and researchers.

We aim to create a new standard reading passage, specifically designed for speech and voice assessment. As, to our knowledge, no guidelines exist for this task, we have initiated this Delphi project in which we would like to invite you to take part. The aim of this project is to involve a large international panel of professionals (clinicians and researchers) active in the fields of speech and voice, in order to reach a consensus on what criteria have to be taken into account when creating a standard reading passage for speech/voice assessment.

Target audience:

This study is addressed to professionals (clinicians, researchers, lecturers) who are currently engaged in activities in at least one of the following fields:

- speech sound disorders (incl. dysarthria, apraxia/dyspraxia, orofacial structural deficits, head and neck oncology, velar insufficiencies, hearing impairment and articulation disorders)

- fluency disorders (stuttering/stammering)

- voice disorders

By 'activities', we understand (if at least approximatively 20% relate to speech/voice):

- clinical activity

- research

- academic activity

- industrial activity

- a combination of the above

The Delphi method:

The Delphi technique is an extensively used group survey methodology that is conducted over several consecutive rounds and aims to reach a consensus among a panel of individuals with expertise (both professional and experience-based) in the investigated field.

This study is quasi-anonymous: the identity of each participant is only known to the main investigator/moderator, and only via the provided email address (to monitor round-to-round response 2

rates); participants remain anonymous to each other, which allows for freedom of expression without any social or professional pressure from peers.

For more information about the Delphi method:

- Diamond IR, Grant RC, Feldman BM, Pencharz PB, Ling SC, Moore AM, et al. Defining consensus: A systematic review recommends methodologic criteria for reporting of Delphi studies. J Clin Epidemiol 2014;67:401–9. https://doi.org/10.1016/j.jclinepi.2013.12.002.

- Hasson F, Keeney S, McKenna H. Research guidelines for the Delphi survey technique. J Adv Nurs 2000;32:1008–15. https://doi.org/10.1046/j.1365-2648.2000.t01-1-01567.x.

- Hsu C-C, Sandford BA. The Delphi technique: Making sense of consensus. Pract Assessment, Res Eval 2007;12. https://doi.org/10.7275/pdz9-th90.

- McPherson S, Reese C, Wendler MC. Methodology update: Delphi studies. Nurs Res 2018;67:404–10. https://doi.org/10.1097/NNR.0000000000000297.

Our study:

As stated above, this Delphi survey yields to seek agreement through an international decision-making process, on what criteria should be taken into account when creating a new reading passage for speech/voice assessment. There also appears to be a lack of consensus regarding the terminology of speech-related concepts as well as the assessment methods, which in turn may influence the decisions taken when implementing new speech assessment materials.

This consensus survey will be structured as follows:

1) Definitions of speech-related concepts*

2) Perceptual and objective speech measures*

3) Criteria for creating standard reading passages

*questions in these sections will only be presented to participants with activities in speech and fluency disorders

In light of the investigated topic and the targeted expert group, we hope to achieve our goal within three (max. four) rounds, between July 2020 and not later than February 2021. Each online survey will be available for 2-3 weeks for you to complete at your convenience and the first round will take about 30-40 minutes to complete (the following rounds will be shorter). None of the surveys will have to be completed in a single sitting, as you will be able to save your answers at any time to resume the questionnaire later. It is very important that participants complete the survey in each round. The reliability of the results could be compromised if participants drop out of the study before its completion. However, it is of course possible at any time to withdraw from the study. To reduce the likelihood of high drop-out rates affecting the outcome of this study, we ask you to only agree to participate if you think you will be available to complete all three to four rounds.

Information notice

Project title: « Delphi consensus survey – Developing reading passages for the assessment of

speech and voice »

Principal investigators responsible for the project: Timothy Pommée (PhD), Julien Pinquier (thesis supervisor), Julie Mauclair and Virginie Woisard (co-supervisors), Renée Speyer Research location: Institut de Recherche en Informatique de Toulouse (Université Paul Sabatier – Toulouse III, France)

Dear Sir or Madam,

We are contacting you in the context of a research project on the development of standard reading passages for speech and voice assessment. This study is part of Timothy Pommée’s PhD thesis at the Institut de Recherche en Informatique de Toulouse (Université Toulouse III Paul Sabatier, France). This project is carried out in collaboration with the Hospitals of Toulouse, the University Cancer Institute of Toulouse Oncopole and the University of Oslo. The primary aim of this thesis is to identify the needs for objective assessment of speech disorders (particularly speech intelligibility), in order to help develop new speech assessment tools that are tailored to the needs of clinicians and researchers.

We also aim to create a new standard reading passage, specifically designed for speech and voice assessment. As, to our knowledge, no guidelines exist for this task, we have initiated this Delphi project in which we would like to invite you to take part. The aim of this project is to involve a large international panel of professionals (clinicians and researchers) active in the fields of speech and voice, in order to reach a consensus on what criteria have to be taken into account when creating a standard reading passage for speech/voice assessment.

Target audience:

This study is addressed to professionals (clinicians, researchers, lecturers) who are currently engaged in activities in at least one of the following fields:

- speech sound disorders (incl. dysarthria, apraxia/dyspraxia, orofacial structural deficits, head and neck oncology, velar insufficiencies, hearing impairment and articulation disorders)

- fluency disorders (stuttering/stammering)

- voice disorders

By 'activities', we understand (if at least approximatively 20% relate to speech/voice):

- clinical activity

- research　

- academic activity

- industrial activity

- a combination of the above

The Delphi method:

The Delphi technique is an extensively used group survey methodology that is conducted over several consecutive rounds and aims to reach a consensus among a panel of individuals with expertise (both professional and experience-based) in the investigated field.

This study is quasi-anonymous: the identity of each participant is only known to the main investigator/moderator, and only via the provided email address (to monitor round-to-round response rates); participants remain anonymous to each other, which allows for freedom of expression without any social or professional pressure from peers.

For more information about the Delphi method:

- Diamond IR, Grant RC, Feldman BM, Pencharz PB, Ling SC, Moore AM, et al. Defining consensus: A systematic review recommends methodologic criteria for reporting of Delphi studies. J Clin Epidemiol 2014;67:401–9. https://doi.org/10.1016/j.jclinepi.2013.12.002.

- Hasson F, Keeney S, McKenna H. Research guidelines for the Delphi survey technique. J Adv Nurs 2000;32:1008–15. https://doi.org/10.1046/j.1365-2648.2000.t01-1-01567.x.

- Hsu C-C, Sandford BA. The Delphi technique: Making sense of consensus. Pract Assessment, Res Eval 2007;12. https://doi.org/10.7275/pdz9-th90.

- McPherson S, Reese C, Wendler MC. Methodology update: Delphi studies. Nurs Res 2018;67:404–10. https://doi.org/10.1097/NNR.0000000000000297.

Our study:

As stated above, this Delphi survey yields to seek agreement through an international decision-making process, on what criteria should be taken into account when creating a new reading passage for speech/voice assessment. There also appears to be a lack of consensus regarding the terminology of speech-related concepts as well as the assessment methods, which in turn may influence the decisions taken when implementing new speech assessment materials.

This consensus survey will be structured as follows:

1) Definitions of speech-related concepts*

2) Perceptual and objective speech measures*

3) Criteria for creating standard reading passages

*questions in these sections will only be presented to participants with activities in speech and fluency disorders

In light of the investigated topic and the targeted expert group, we hope to achieve our goal within three (max. four) rounds, between July 2020 and not later than February 2021. Each online survey will be available for 2-3 weeks for you to complete at your convenience and the first round will take about 30-40 minutes to complete (the following rounds will be shorter). None of the surveys will have to be completed in a single sitting, as you will be able to save your answers at any time to resume the questionnaire later. It is very important that participants complete the survey in each round. The

reliability of the results could be compromised if participants drop out of the study before its completion. However, it is of course possible at any time to withdraw from the study. To reduce the likelihood of high drop-out rates affecting the outcome of this study, we ask you to only agree to participate if you think you will be available to complete all three to four rounds.

Top

Dear All,

Top