Abraham Owodunni

I am a Ph.D. student in Computer Science and Engineering at The Ohio State University where I have been privilaged be advised by the awesome Prof. Sachin Kumar. Before that, I was a researcher at IntronHealth during which I work on creating language and speech technologies to African accents and languages (including MENA).

I am also an independent researcher at Masakhane, an open science community where Africans do AI research for Africa. I am primarily interested in investigating novel ways to develop robust multilingual models in resource-constrained settings, particularly using efficient ML techniques.

Beyond the technical aspect of research, I worked on creating national AI policies within Nigerian and Africa, I enjoy mentoring new folks that are interested in getting into AI research and I love building and nurturing open science ML communities. You’ll find me around Sisonke Biotik, ML Collective, Cohere4AI and a few others. I firmly believe in democratizing access to knowledge, fostering collaborative ecosystems, and championing the ethos of shared discovery.

If I’m not doing AI research, you can find me playing Pool, Lawn and Table Tennis (I suck at both though), reading books and listening podcasts.

Podcast recommendations:

How To Take Over The World

news

Sep 2, 2024	I served at the Publication chair for the MRL 2024 Workshop at EMNLP. Checkout our proceedings.
Aug 16, 2024	I resumed my Ph.D. in Computer Science and Engineering at OSU.
Jun 4, 2024	2 papers on TTS and STT got accetped to Interspeech 2024! Find them here [1, 2].
Jan 18, 2024	Our papers `AccentFold: A Journey through African Accents for Zero-Shot ASR Adaptation to Target Accents` got accetped to EACL Finding 2024!
Nov 25, 2023	I gave a talk at DevFest Abuja titled “Growing AI Talents with Open Science Communities”. Slides

latest posts

selected publications

Bloom Library: Multimodal Datasets in 300+ Languages for a Variety of Downstream Tasks

Colin Leong, Joshua Nemecek, Jacob Mansdorfer, and 3 more authors

In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, Dec 2022

Abs

We present Bloom Library, a linguistically diverse set of multimodal and multilingual datasets for language modeling, image captioning, visual storytelling, and speech synthesis/recognition. These datasets represent either the most, or among the most, multilingual datasets for each of the included downstream tasks. In total, the initial release of the Bloom Library datasets covers 363 languages across 32 language families. We train downstream task models for various languages represented in the data, showing the viability of the data for future work in low-resource, multimodal NLP and establishing the first known baselines for these downstream tasks in certain languages (e.g., Bisu [bzi], with an estimated population of 700 users). Some of these first-of-their-kind baselines are comparable to state-of-the-art performance for higher-resourced languages. The Bloom Library datasets are released under Creative Commons licenses on the Hugging Face datasets hub to catalyze more linguistically diverse research in the included downstream tasks.
ICLR: AfricaNLP

Koya: A Recommender System for Large Language Model Selection

Abraham Toluwase Owodunni, and Chris Chinenye Emezue

In 4th Workshop on African Natural Language Processing, Dec 2023