2023-09-27
Language technologies are inherent in human daily lives. Recent advances in AI and Natural Language Processing technologies showed how powerful computers could be useful to humans, especially as a means of communication. Unfortunately, existing systems are biased (gender, and race), and barely integrate low-resourced languages; while becoming more and more computationally expensive.
In this presentation, as a part of his work, Bonaventure will share with us his experience working on methods leveraging how we can build efficient language technologies without relying on huge amounts of data, and computation, while also mitigating bias in language representation.
Bonaventure Dossou, MSc Student in Data Engineering, Jacobs University Bremen and MSc Research Intern, Mila
Bonaventure F. P. Dossou is pursuing his doctorate at the Center for Research on Intelligent Machines, McGill University. He holds a Bachelor of Science in Mathematics (Russia) and a Master of Science in Computer Science and Data Engineering (Germany). Previously, he worked as a researcher at the Mila Quebec AI Institute, Google Research, Roche Canada, and Modelis, to name but a few. Bonaventure is interested in natural language processing for African languages and machine learning for healthcare. He is actively working on the creation of linguistic datasets and technologies such as multilingual machine translation systems (like FFR, MMTAfrica), multilingual speech recognition systems (like Okwugbe), and large language models (like AfroLM) for African languages.