Starting date: February or March 2024 (flexible)

Place of work:
Laboratoire d'Informatique de Grenoble, CNRS, Grenoble, France

Duration: 6 months

Keywords: natural language processing, speech processing, dependency
parsing, treebanks

Contact: maximin.coavoux@univ-grenoble-alpes.fr

Description:

The SynPaX ANR project (Syntactic Parsing of Spontaneous speech with
cross-modal transfer learning) aims at investigating modal fusions
systems to develop end-to-end speech parsing models, i.e. systems that
work directly on speech signal to perform parsing, instead of using
only (potentially noisy) speech transcriptions as input.

The two main bottlenecks in parsing speech are: (i) the scarcity of the
available training data (ii) the difficulty in transcribing spontaneous
speech. The internship will focus on (i) and explore data augmentation
methods by leveraging both audio corpora without syntactic annotations
and existing treebanks for written texts.


Tasks:

-   literature review on speech parsing
-   designing and implementing data augmentation methods for speech
    parsing
-   empirical evaluation of proposed methods


Scientific environment:

The internship will be conducted within the Getalp team of the LIG
laboratory (https://lig-getalp.imag.fr/). The GETALP team has a strong
expertise and track record in Natural Language Processing. The
recruited person will be welcomed within the team which offer a
stimulating, multinational and pleasant working environment.
The internship will be supervised by Maximin Coavoux, Adrien Pupier
and Benjamin Lecouteux.

The SynPaX ANR project will also fund a PhD scholarship starting in
Fall 2024 to work on extensions of this internship.


Profile:

-   be enrolled in a Master (M2) in NLP, computational linguistics or
    computer science
-   background in NLP and/or speech processing
-   proficiency in Python (transformers, pytorch)
-   Good communication skills in English, rudimentary knowledge of
    French is also expected since the internship will focus on French
    data.


How to apply:

-   please send CV + cover letter + recent (last ~2 years) academic
    transcripts to maximin.coavoux@univ-grenoble-alpes.fr


References:

End-to-End Dependency Parsing of Spoken French (Interspeech 2022)
Adrien Pupier, Maximin Coavoux, Benjamin Lecouteux, Jérôme Goulian
https://hal.science/hal-03713551/

Wave to Syntax: Probing spoken language models for syntax
Gaofei Shen, Afra Alishahi, Arianna Bisazza, Grzegorz Chrupala
https://arxiv.org/abs/2305.18957