Job: Postdoc at UCLouvain (1-3 years): automatic assessment of learner’s productions in FFL

[Adrien Bibal]

The UCLouvain is looking for a

Postdoctoral researcher in Machine Learning or Natural Language Processing

• Fixed-term full-time (100%) contract for one year, renewable twice

• At the Center for Natural Language Processing (Centre de traitement automatique du langage; Cental), at the Institute for Language and Communication (ILC) at the UCLouvain (Louvain-la-Neuve)

• Starting date: September 2021

This postdoctoral research position is integrated within an ambitious research project conducted at the Center for Natural Language Processing (Centre de traitement automatique du langage; Cental; https://uclouvain.be/fr/instituts-recherche/ilc/cental), in collaboration with FEI (France Éducation International; https://www.france-education-international.fr). Funding is available for 3 years, but employment takes the form of one-year fixed-term contracts. Renewal of the contract is contingent upon the realization of research objectives defined within the project.

The project will bring together multiple researchers to develop an automated solution to assist and support the correction of writing skill tests for the French language for certificate evaluation purposes. Specifically, the project aims at developing a system that can (1) automatically evaluate the CEFR level of writing skills of test takers by training an artificial intelligence system on data provided by FEI, and (2) provide various diagnostic indicators based on pedagogical resources in French as a Foreign Language. The first objective is in line with work on English aimed at the automatic evaluation of learner productions in relation to the Common European Framework of Reference (CEFR) (Tack et al., 2017; Arnold et al., 2018; Yannakoudakis et al., 2018; Baillier et al., 2019). The second objective is in line with work conducted at the Cental on the diagnosis of complexity of written productions (CEFRLex in François et al., 2014; Pintard and François, 2020 or AMesure in François et al., 2020).

The tasks of the postdoctoral researcher will mainly consist in (1) preparing the FEI corpus for statistical treatment, (2) implementing a Python package for the extraction of linguistic (e.g. measuring the complexity of lexical or syntactical structures, cf. Tack et al., 2017) and pedagogical features (e.g. developmental stages, cf. Bartning and Schlyter, 2004) correlated with the CEFR level of productions of learners of French as a Foreign Language, (3) developing an AI model capable of automatically predicting this CEFR level, and (4) testing the system with human assessors. The postdoctoral researcher will also be responsible for the dissemination of results via scientific publications and yearly reports. The researcher will work in close collaboration with the other team members to ensure the successful implementation of these tasks.

Work environment

The Cental is part of the Institute for Language & Communication (https://uclouvain.be/fr/instituts-recherche/ilc) which is part of the University of Louvain. The university is situated in Louvain-la-Neuve (https://uclouvain.be/fr/sites/louvain-la-neuve), a pleasant and dynamic pedestrian city. The research project is headed by Prof. Thomas François (https://cental.uclouvain.be/team/tfrancois/), an expert in readability and automatic text simplification. Research missions to Paris (Sèvres) will be organized to ensure and maintain close collaboration with FEI. FEI is internationally known for its expertise in the domain of proficiency level evaluation for French as a Foreign Language and is responsible for the organization of the main French language tests, namely TCF, DELF and DALF.

Qualifications

The ideal candidate will have:

• A PhD in Information Technology, Computational Linguistics, Natural Language Processing or equivalent

• Excellent computer skills:

  • Programming languages : Python (mastery)

  • Experience with sckit-learn, pandas, TensorFlow, PyTorch or Keras is a plus

  • Web technologies: HTML, JavaScript, CSS, Flask, MySQL and PostGreSQL

  • Operating systems: Linux (server management)

  • Experience in the creation and dissemination of Python libraries via Github (or equivalent) is a plus

• Knowledge of the main algorithms used in supervised machine learning is required. Knowledge of Deep Learning is a plus

• Excellent knowledge of English (at least level C1) and good knowledge of French (at least level B2)

• Excellent research profile (publications, conferences, etc.)

• Firm grasp of the main tools and algorithms in NLP

• Research experience in one of the following domains is a plus: computer-assisted language learning, readability, automated essay scoring, language assessment, etc.

• Autonomy, team spirit, capacity to listen and analyze needs, reactivity

Employment conditions

This fixed-term postdoctoral position is subject to following conditions:

• Fixed-term one-year contract, renewable twice, for a possible total of three years

• Gross salary is commensurate with seniority and will vary between 4250€ and 4850€ per month

This position normally requires a relocation to Belgium. Applicants from outside the EU are responsible for obtaining a visa and required permits, with the support of the corresponding department at the UCLouvain.

Application

Application deadline: 30 June 2021

If you are interested in this position, please send your application to Thomas François (thomas....@uclouvain.be). The application must include:

1. A detailed CV in French or English including qualifications and required skills, publications and other academic and scientific experience

2. A letter of motivation in French describing your interest for the position, how your profile corresponds to the description of the position and to the objectives of the project, etc.

3. A letter of reference in French or English

Short-listed candidates will be invited to participate in a video conference interview. The exact modalities will be communicated via mail.