2022 International Conference on Applied Mathematics, Informatics, and Computing Sciences (AMICS 2022)

Abstract: The abundant availability of data in Big Data era has helped achieving significant advances in the machine learning field. However, many high-dimensional datasets appear with incompleteness from different perspectives such as values, labels, annotations and records. Consequently, many machine learning methods struggle in providing reliable predictive and descriptive models. The implemented algorithms often struggle with learning a generalized model. The curse of dimensionality concept encompasses these challenges. Bayesian networks offer a promising remedy to these challenges as they rely on a strong graphical and probabilistic foundation. Their learning algorithms are robust against the noisy and incomplete data. Their underlying models are therefore capable of ensuring generalization even when learnt from high-dimensional datasets having small volumes. In this talk, these challenges will be discussed and the Bayesian network foundations will be explained. Additionally, the emphasis is put on the merits of Bayesian networks in overcoming the curse of dimensionality.