ILES's research topics are organised with the following four topics:
- Corpora and representations
Defining this information, its association with language, and how to evaluate automated systems which aim to process language in a meaningful way with respect to this information, is the goal of the Corpora and Representations topic. This is a foundational activity with a direct, concrete impact on the definition, training, and evaluation of language processing systems. ILES has produced and made available a number of annotation models, guidelines, annotated corpora, and evaluation metrics, and organized shared tasks to evaluate natural language processing systems based on these corpora.
- Multilingualism and Paraphrasing
Determining which variations in language form do or do not preserve the same information, within one language or across multiple languages, investigates a characteristic property of natural languages: the non-bijection between form and meaning. This is the topic addressed by Multilingualism and Paraphrasing.
- Information Extraction, Focused Information Retrieval, and Question Answering
Designing methods to extract and query information from natural language input gives access to otherwise unattainable information and knowledge.
- Sign language Modelling and Processing
The goals are to produce of linguistic resources and to tackle problems of analysis, representation and processing of Sign Language in an interdisciplinary way, with points of view from several fields of informatics (NLP, signal processing, computer graphics), as well as the sciences of language, movement and perception.