I am a Machine Learning Researcher and Data Scientist. Currently, I am a Senior Lecturer (Assistant Professor) in AI at the Victoria University of Wellington (VuW), New Zealand. Before joining VuW, I worked at the University of Waikato as a Senior Researcher and Co-Director of the AI Institute. My main research area is on Adaptive Machine Learning, which includes machine learning for data streams and, more recently, online continual learning. For collaborations and inquires about PhD positions reach out through my institutional email.

CapyMOA

I lead the development of CapyMOA, a machine learning library tailored for data streams. CapyMOA features a Python API that leverages MOA’s efficient implementations, PyTorch for Online Continual Learning and neural networks support, and scikit-learn for integration with classical incremental learning algorithms.

• Website: https://capymoa.org/
• CapyMOA Github: https://github.com/adaptive-machine-learning/CapyMOA
• We had plenty of tutorials in 2024-2025, including: PAKDD (Taipei, Taiwan), IJCAI (Jeju, South Korea), KDD (Barcelona, Spain), ECML (Vilnius, Lithuania), KiwiPycon (Wellington, NZ) and ICONIP (Auckland, NZ), ICDE (Hong Kong, SAR China), PAKDD (Sydney, Australia), PRICAI (Wellington, NZ) and more. Material is available on the CapyMOA discord here

Selected Publications

Kolmogorov-Arnold Networks Still Catastrophically Forget but Differently from MLP

A Lee, H M Gomes, Y Zhang, W B Kleijn. AAAI Conference on Artificial Intelligence, 2025 Paper

We give a simple explanation as to why and how KAN catastrophically forget. Motivated by evidence suggesting KAN are superior for symbolic regression, we augment KAN in the same ways as multilayer perceptron (MLP) to perform continual learning tasks, making special accommodations to support KAN.

Leveraging Plasticity in Incremental Decision Trees

M Heyden, H M Gomes, E Fouché, B Pfahringer, K Böhm European Conference on Machine Learning (ECML PKDD), 2024. Paper

PLASTIC is a novel incremental decision tree that yields better predictive performance than other recent algorithms, such as the canonical Extremely Fast Decision Tree (EFDT). PLASTIC restructures the otherwise pruned subtree. This is possible due to decision tree plasticity: one can alter a tree’s structure without affecting its predictions.

Gradient boosted trees for evolving data streams

N Gunasekara, B Pfahringer, H M Gomes, A Bifet. Machine Learning, Springer, 2024. Access paper

Streaming Gradient Boosted Trees (SGBT) is trained using weighted squared loss elicited in XGBoost. SGBT employs a tree replacement strategy that detects and recovers from drifts, thus enabling the ensemble to adapt without sacrificing the predictive performance.

Machine learning (in) security: A stream of problems

F Ceschin, M Botacin, A Bifet, B Pfahringer, L S Oliveira, H M Gomes, A Gregio ACM Digital Threats: Research and Practice, 2024 Paper

We identify, detail, and discuss the main challenges in the correct application of Stream ML techniques to cybersecurity data. We evaluate how concept drift, evolution, delayed labels, and adversarial ML impact the existing solutions.

Survey on Online Streaming Continual Learning.

N Gunasekara, B Pfahringer, HM Gomes, A Bifet. International Joint Conference on Artificial Intelligence (IJCAI), 2023. Paper

This survey explores the intersection of Continual Learning (CL) and Streaming Learning (SL). This papers includes a gentle introduction to SL and CL, then we point out new research trends and directions for future research.

A Survey on Semi-Supervised Learning for Delayed Partially Labelled Data Streams

H M Gomes, M Grzenda, R Mello, J Read, M H L Nguyen, A Bifet. ACM Computing Surveys, 2022. DOI: https://doi.org/10.1145/3523055

We organise methods that leverage unlabelled data in a semi-supervised setting for streaming data. We also discuss the delayed labelling issue, which impacts both fully supervised and semi-supervised methods.

Streaming Random Patches for Evolving Data Stream Classification

H M Gomes, J Read, A Bifet. IEEE International Conference on Data Mining (ICDM), 2019. DOI: https://doi.org/10.1109/ICDM.2019.00034

The Streaming Random Patches (SRP) algorithm outperforms the current state-of-the-art ensemble methods for evolving data stream classification. Paper

Machine learning for streaming data: state of the art, challenges, and opportunities

H M Gomes, J Read, A Bifet, J P Barddal, J Gama. SIGKDD Explorations Newsletter, ACM , 2019. DOI: https://doi.org/10.1145/3373464.3373470

In this work, we focus on elucidating the connections among the current stateof-the-art on related fields; and clarifying open challenges in both academia and industry. Paper

A Survey on Ensemble Learning for Data Stream Classification

H M Gomes, J P Barddal, F Enembreck, A Bifet. ACM Computing Surveys 50, 2, Article 23, 2017. DOI: https://doi.org/10.1145/3054925

This paper contains a comprehensive survey about ensemble learning for data streams. Paper

Adaptive random forests for evolving data stream classification

H M Gomes, A Bifet, J Read, …, B Pfahringer, G Holmes, T Abdessalem. Machine Learning, Springer, 2017. DOI: https://doi.org/10.1007/s10994-017-5642-8

This paper presents an efficent version of the classical Random Forests algorithm for evolving data streams, namely the Adaptive Random Forest (ARF) algorithm. Paper