Maties Machine Learning (MML) is a seminar series and discussion forum with the goal of bringing together people working on machine learning at Stellenbosch University. We meet roughly every second week for short talks on people’s current work, some ML-related topic, or open discussions. The idea is to get to know what others are working on and to strengthen machine learning research at Stellenbosch.

Upcoming talks

---

15 May 2026, 13:10-14:00 in room A3008, General Engineering Building

Batsi Ziki – An introduction to meta-continual learning

For real world applications it is ideal if machine learning models have the ability to continually learn new tasks; in reality, it is challenging for machine learning models to learn tasks sequentially due to catastrophic forgetting. Continual learning is a field concerned with designing algorithms that mitigate catastrophic forgetting; however, these algorithms are hand-designed. Researchers have looked at combining continual learning and meta-learning to design general continual learning algorithms. Meta-continual learning is the result of this combination and is informally defined as learning how to continually learn. In this talk, we will walk through meta-continual learning by introducing continual learning and meta-learning, before finishing off with Generative Meta-Continual Learning (GeMCL), an algorithm that is immune to catastrophic forgetting.

Short bio: Batsi is a research scientist at ByteFuse.ai

---