We continue to join forces with Qulinda to discuss deep learning, wildlife and livestock. Identifying and tracking individuals in vast nature reserves is crucial for supporting wildlife biodiversity and welfare and addressing conflicts at t…
Selecting a research topic is a daunting task. Here are some things to help you think through the process. The first challenge to beginning your research study is finding a topic you are interested in. Then, you must find a niche within tha…
Machine learning has so many words and ideas that it’s becoming increasingly difficult to know what is happening. I have tried to capture some coherence here to aid discussions. These are constraints placed on the learning due to the availa…
Ideally machine learning models should be capable of learning without “explicit” supervison and do so continously without forgetting what they have previously learned . However, neither of these two problems are satisfactorily solved in dee…
We recently joined forces with Smart Savanahs to bring deep learning models into Sub-Saharan Africa. Identifying and tracking individuals in vast nature reserves is crucial for supporting wildlife biodiversity and welfare and addressing con…
The Applied Representation Learning Lab as part of The Institute for Intelligent Systems (IIS) is now a research group node under Centre for Artificial Intelligence Research (CAIR) Representation learning underpins forecasting, prediction a…
Agent-Based Models (ABMs) as complex adaptive simulations which are able to incorporate real world nuance into their design. This research looks at the surrogate (machine learning) assisted agent based modelling. Agent-Based Models (ABMs) a…
The recent paper by Variawa and van Zyl inter al. (2022) [ref] looks at transfer learning and deep metric learning for galaxy morphology classification. The most exciting result from this paper is that metric learning outperformed tradition…
This research project looks at the impact of catastrophic forgetting and incremental learning in similarity learning. Incremental learning is when we continuously update a machine learning model with new objectives (classes or tasks). The a…
This research project looks at the intersection of finance and machine learning. Online Portfolio Selection is a fundamental problem in Computational Finance. Our recent pattern-Matching method, DRICORN-K: A Dynamic RIsk CORrelation-driven
Interesting applications arises at the intersection of evolutionary computing and machine learning when we capitalise on the strengths of both of these paradigms simultaneously. Computational Intelligence, specifically evolutionary computin…
Studied as a Computer Scientist in Artificial Intelligence and Complex Adaptive Systems. Working as a Professor in Machine Learning at the University of Johannesburg. Currently, hold the Nedbank research and innovation chair. Also an NVIDIA…