My current research interests lie in:
Weakly supervised machine learning approaches for efficient data aquisition and annotation, especially for scientific machine learning.
Machine learning for combinatorial problems, which I'm exploring on coalition structure generation problems.
My doctoral thesis (available here) focused on center-based clustering methods, that operate in an unsupervised or weakly supervised manner.
Multiple Kernel Transfer Clustering (Github): In a source task a multiple kernel metric is learnt under weak supervision, which is transferred to a target task where a data set is clustered.
Entropy c-Means Clustering (Github): A multi-objective fuzzy clustering method to identify clusters at different levels of overlap.
The Last Leap and the Last Major Leap (Github): Methods to estimate the number of clusters in a data set.
Kernel k-Harmonic Means Clustering (Github): An extension of the k-Harmonic Means method incorporating kernel similarities.
