alternative

Prof. Juraj Šimko

University of Helsinki
Finland

Prior to joining the University of Helsinki in 2013, Dr. Šimko have studied and worked at several Universities in Slovakia (MSc in Mathematics), Ireland (PhD in Computer Science) and Germany (von Humboldt Fellowship), and spent several years as a Language Specialist in Microsoft. His main areas of expertise are modelling of speech articulation, prosody, speech synthesis and using large deep learning-based models to answer theoretical questions related to speech.

alternative

Prof. Priyankoo Sarmah

Indian Institute of Technology Guwahati
India

Priyankoo Sarmah is a Professor of linguistics at the Indian Institute of Technology Guwahati. He has a PhD in linguistics from the University of Florida. He has been working on the phonetics of tones and vowels of several North East Indian languages. Over the years, he has worked on speech technology development (speech recognition, dialect classification and text to speech synthesis) for resource-poor languages.

  • Course Overview

    • Phonetics and technology

    This course is designed to bridge the pedagogic gap between phonetics and technology. In this version of the course, the instructors will train the participants in prosody analysis, both in traditional manner and using machine learning techniques. The analyses will be applied to address typological hypotheses concerning language contact, language family relations, sound change, etc.

    Background

    Linguistic diversity in India is immense. There are, however, relatively few studies on the linguistic and prosodic characteristics that result from interactions among multiple languages widely used by diverse communities. One of the difficulties facing multilingual research in India is a shortage of annotated, compatible speech data available for the local languages. Nevertheless, the recent breakthroughs in speech technology and machine learning allow for novel ways of investigating consequences of multilingualism on ever-changing typological properties of different languages even without annotated speech databases.

    Application

    We will present and discuss the recently developed techniques of phonetic, prosodic and typological analysis that use computational power of deep learning to by-pass the need of manual corpus annotations, and can be applied for investigation of existing corpora that contains speech data capturing linguistic, dialectal or social variation of speakers.

    Who is this course for?:

    • Students of linguistics and phonetics who want to learn more about prosodic analysis and machine learning
    • Students of technology and engineering that want to find out about the vast richness of research in linguistics, phonetics and typology
    • Early career researchers with passion in linguistics, dialectology, or sociolinguistics, and want to learn about the state of the art ways of analysing
    Schedule
    January 20, Monday L1: Introduction to linguistic variation and prosodic typology
    L2: Prosodic/phonetic features
    January 21, Tuesday L3: Examples and reviews of machine learning based prosodic analyses
    L4: Corpus structure
    T1: Examples of speech corpora
    January 22, Wednesday L5: Classifiers: supervised, semisupervised and selfsupervised learning
    T2: Hands on work with classifiers
    L6: The use of classifiers for addressing research questions in prosodic typology
    January 23, Thursday L7: Speech technology and prosodic variation: latent spaces, embeddings
    T3: Hands on work with latent prosodic representations
    L8: Use of latent representations for prosodic typology
    January 24, Friday L9: Challenges and opportunities of digital prosodic typology in the diverse language environment of India

    Registration Procedure

    1. Candidate should apply to the course on or before Dec 22, 2024 using the Application form link:
    2. After shortlisting the candidate will be notified by January 3, 2025 via email.
    3. Shortlisted candidate should pay the registration fee and provide the details in the below form by January 12, 2025:





    Course fees:
    1. Industry/ Organizations: Rs. 9000/- (including GST)
    2. Non-students: Rs. 5000/- (including GST)
    3. Students: Rs. 1000/- (including GST)
    1. The selected candidates must pay the applicable fees by online bank transfer / wire transfer / internet banking to the following bank account. Please keep the online transfer receipt for proof of transfer.
    2. Mode of participation will be both offline and online. About 70 seats are available for offline participation.
    3. For offline participants accommodation will be provided on payment basis.
    Paymemt Details

    Bank Name : STATE BANK OF INDIA
    Branch Name : IIT GUWAHATI BRANCH
    IFSC Code : SBIN0014262
    MICR code : 781002053
    Account Name : IIT GUWAHATI R&D – MHRD
    Account No : 31151533220
    Account Type : Savings
     

    Prof. Priyankoo Sarmah

    Professor
    Department of Humanities & Social Sciences, IIT Guwahati
    Guwahati -781039
    Phone: +91 361 258-2574
    E-mail:priyankoo@iitg.ac.in
    Helpful Links: