Mahi Aminu Aliyu

Bauchi, Nigeria

Doumo(“Hey” in Japanese)!

Welcome to my personal page.

I am an independent ML researcher and a second-year MSc student at Abubakar Tafawa Balewa University. I completed my degree in Computer Science at Al-Qalam University Katsina in 2023. My research focuses on causal representation learning and inductive biases — specifically, how incorporating causal structure into learned representations can improve robustness and out-of-distribution generalization, with applications in low-resource NLP and public health.

news

Apr 20, 2026	Exciting News: I’m excited to share that I’ve been accepted into the Fatima Fellowship 2026!
Apr 01, 2026	Exciting News: I left my role at AHFID to join Child Health and Mortality Surveillance (CHAMPS) Health - ATBU as a full-time Analyst/Programmer. Excited to work with longitudinal data.
Jan 26, 2026	Exciting News: Our work on AFRICAPTION: Establishing a New Paradigm for Image Captioning in African Languages has been accepted into the 7th AfricaNLP workshop co-located with EACL. arXiv:2510.17405
Nov 20, 2025	Our preprint “When Distributions Shift: Causal Generalization for Low-Resource Languages” is now available on arXiv. arXiv:2510.27512
Oct 15, 2025	I joined Africa Hub for Innovation and Development (AHFID) as Artificial Intelligence and Social Informatics Associate
Sep 14, 2025	Exciting News: My preliminary work on water quality estimation using machine learning, “Analysis of Machine Learning Models for Efficient Water Quality Prediction,” has been accepted in the Journal of Information Systems and Technology Management (JISTM).
Aug 10, 2025	I will be attending Deep Learning Indaba 2025 in Kigali. #URUNANA
Jun 30, 2025	Exciting News: I’ve been accepted to attend the Machine Learning Summer School — Senegal 2025 !
Jan 27, 2025	I gave a talk on my ongoing research project at ML Collective research jam #25 titled “A pre trained model for water-based applications”, you can watch the talk here
Jul 01, 2024	I am now the City Lead Data Science Nigeria Bauchi chapter.
Mar 18, 2024	I joined ML Collective a non-profit independent research organisation! I am heading the ML Theory and Optimisation research group with a focus on representation learning.
Jan 01, 2024	I am currently pursuing an MSc in Computer Science at Abubakar Tafawa Balewa University
Apr 25, 2023	I bagged my BSc degree in Computer Science from Al-Qalam University Katsina, YAY!
Jan 01, 2023	I have joined Horizon IT Services as a data analyst.

selected publications

arXiv
When Distributions Shift: Causal Generalization for Low-Resource Languages

Mahi Aminu Aliyu , Chisom Chibuike , Fatimo Adebanjo , and 2 more authors

arXiv preprint arXiv:2510.27512, 2025

Abs arXiv Bib

Machine learning models often fail under distribution shifts, a problem exacerbated in low-resource settings where limited data restricts robust generalization. Domain generalization (DG) methods address this challenge by learning representations that remain invariant across domains, frequently leveraging causal principles. In this work, we study two causal DG approaches for low-resource natural language processing. First, we apply causal data augmentation using GPT-4o-mini to generate counterfactual paraphrases for sentiment classification on the NaijaSenti Twitter corpus in Yoruba and Igbo. Second, we investigate invariant causal representation learning with the Debiasing in Aspect Review (DINER) framework for aspect-based sentiment analysis. We extend DINER to a multilingual setting by introducing Afri-SemEval, a dataset of 17 languages translated from SemEval-2014 Task. Experiments show improved robustness to unseen domains, with consistent gains from counterfactual augmentation and enhanced out-of-distribution performance from causal representation learning across multiple languages.
@article{aliyu2025causal, title = {When Distributions Shift: Causal Generalization for Low-Resource Languages}, author = {Aliyu, Mahi Aminu and Chibuike, Chisom and Adebanjo, Fatimo and Awosanya, Omokolade and Oyeneye, Samuel}, journal = {arXiv preprint arXiv:2510.27512}, year = {2025}, }
AfricaNLP
AfriCaption: Establishing a New Paradigm for Image Captioning in African Languages

Mardiyyah Oduwole , Prince Mireku , Fatimo Adebanjo , and 3 more authors

In Proceedings of the 7th Workshop on African Natural Language Processing (AfricaNLP), co-located with EACL 2026 , 2026

Abs arXiv Bib

Multimodal AI research has overwhelmingly focused on high-resource languages, hindering the democratization of advancements in the field. To address this, we present AfriCaption, a comprehensive framework for multilingual image captioning in 20 African languages and our contributions are threefold: (i) a curated dataset built on Flickr8k, featuring semantically aligned captions generated via a context-aware selection and translation process; (ii) a dynamic, context-preserving pipeline that ensures ongoing quality through model ensembling and adaptive substitution; and (iii) the AfriCaption model, a 0.5B parameter vision-to-text architecture that integrates SigLIP and NLLB200 for caption generation across under-represented languages. This unified framework ensures ongoing data quality and establishes the first scalable image-captioning resource for under-represented African languages, laying the groundwork for truly inclusive multimodal AI.
@inproceedings{oduwole2026africaption, title = {AfriCaption: Establishing a New Paradigm for Image Captioning in African Languages}, author = {Oduwole, Mardiyyah and Mireku, Prince and Adebanjo, Fatimo and Olajide, Oluwatosin and Aliyu, Mahi Aminu and Novikova, Jekaterina}, booktitle = {Proceedings of the 7th Workshop on African Natural Language Processing (AfricaNLP), co-located with EACL 2026}, year = {2026}, }