I am a third year PhD student at the UCL FAI CDT and ELLIS, supervised by Matt Kusner and Michael Bronstein. I also do Bayesian Optimisation research at Matterhorn Studio.
Previously I was a MSc student at the African Masters of Machine Intelligence, and a research intern at the Vector Institute with Rahul G. Krishnan working on uncertainty measurement for generalization. I obtained my BSc in Computer Science from Ashesi University in 2020.
I am broadly interested in modeling techniques that use the (sometimes implicit) structures present in data generating processes, datasets, and models’ internal representations. More specifically, at the moment, I am interested in understanding and improving message passing models, domain generalisation and adaptation, and multi-fidelity Bayesian optimisation
Preprints
- Setting the Record Straight on Transformer Oversmoothing
- Gbètondji J-S Dovonon, Michael Bronstein, Matt J. Kusner
- paper
- Spawrious: A Benchmark for Fine Control of Spurious Correlation Biases
Papers
- Long-run Behaviour of Multi-fidelity Bayesian Optimisation
- Gbètondji J-S Dovonon, Jakob Zeitler
- NeurIPS 2023 Workshop on Adaptive Experimental Design and Active Learning in the Real World, 2023
- paper
- Towards a better understanding of the computer vision research community in africa
- Abdul Hakeem Omotayo*, Mai Gamal*, Eman Ehab*, Gbètondji J-S Dovonon*, Zainab Akinjobi*, Ismaila Lukman*, Houcemeddine Turki*, Mahmoud Abdien*, Idriss Tondji*, Abigail Oppong*, Yvan Pimi*, Karim Gamal*, Mennatullah Siam
- EAAMO: Equity and Access in Algorithms, Mechanisms, 2023
- paper