Linear SVMs for NLP : TF-IDF, Hyperplans et Marges
Introduction
{# Title and description will be output by jekyll-seo-tag to avoid duplicates #}
Photo taken in Serena Hotel, Lake Kivu, Rwanda
I'm Jérémie Mabiala. I studied mathematics and somewhere along the way got curious about what machines can learn.
I'm an MSc student in the African Master in Machine Intelligence (AMMI) program at AIMS Senegal. Before that, I did a Master's in Mathematical Sciences at Stellenbosch University & AIMS South Africa. Still learning — ML theory, NLP, computer vision, LLMs, and optimization interest me most, but I'm open to almost anything AI & math-related.
Before this, I ran practical sessions at the Department of Mathematics and Computer Science, University of Kinshasa. I now help run Python, PyTorch, and Machine Learning bootcamps at AIMS Senegal — teaching remains one of the best ways I know to actually understand something.
Outside of that, I freelance in software development, and I read novels.
Awarded the fully funded AIMS scholarship for academic excellence.
Ran bootcamps in Advanced Linear Algebra, Python, and PyTorch.
Taught undergraduate courses in Real Analysis and Linear Algebra. Graded assignments and supported students in their coursework.
Our paper “Interpretable Property Insurance Premium Prediction Using Machine Learning Models” has been accepted and p...... more
I was awarded a fully funded AMMI resident tutor fellowship to teach in the Machine Learning stream at the African In...... more
I was awarded a fully funded scholarship to study Mathematical Sciences at AIMS South Africa, affiliated with Stellen...... more
A small mobile app developed with React Native and OpenAI APIs to generate human-like images.
AUBAINE, PG2J+X2M
Dakar, Senegal
Did you know? A linear SVM classifier is really just a hyperplane splitting your text vectors — and the margin is exactly a distance formula in disguise.
Read the post