Final Appraisal Seminar for Tenure Track Assistant Professor Desmond Elliott

Portrait of Desmond

Program

16.25-16.30: Short welcome and introduction by the Associate Dean for Research

16.30-17.15 : 45 min presentation by the candidate consisting of the following:  

    • Presentation of the research conducted during the tenure track period and its relation to the state of the art (about 30 minutes)  
    • Presentation of future research visions (about 7-8 minutes)  
    • Reflections about teaching experience and presentation of visions for future teaching development (about 7-8 minutes)    

17.15 – 17.20: Short break  

17.20 – 17.40: Research questions by external opponent, Professor Luke Zettlemoyer

17.40 – 18.00: Research questions by external opponent, Professor Marie-Francine Moens

18.00 – 18.15: Questions to teaching experience and visions by deputy head for education, Ken Friis Larsen

18.15 – 18.30: Questions regarding future ideas and visions, collaboration, international networking etc. by head of department and associate dean for research  

18.30 – 18.45: Evaluation by Tenure Track appraisal committee (closed session - NB – in zoom breakout room - candidate and the audience stay in main room)

18.45 – 19.00: Concluding evaluation and conversation between candidate and tenure track appraisal committee in front of the whole audience.

19:00 – 20:00: Reception

Registration

You only need to register if you participate physically. 

Title

Multilingual and Multimodal Language Processing for All Languages

Abstract

My research focuses on the design and evaluation of multimodal machine learning models that process inputs from different input modalities, such as images, text, and speech. I have studied these models from the perspective of general-purpose representation learning, and in downstream tasks like automatic image captioning or cross-modal retrieval. Most prior work in this area has focused on the English language, due to the limited availability of datasets. In contrast, my research has focused on the importance of multilingual processing, through the collection of non-English datasets in a culturally-engaged process. I will present approaches to collecting and modelling culturally-relevant multimodal data. A second strand of my research focuses on tokenization-free language modelling using visual text representations, directly connecting my interests in multimodal and multilingual learning. I will also present my recent work on creating language models that can process text in any written language. I will conclude by presenting an overview of my teaching and service contributions to the department, and my future plans for research and teaching.