PUBLIC VIVA-VOCE NOTIFICATION
11:00 am, Seminar Hall
Some Computational Approaches for Machine Learning on Big Datasets
Rajiv Sambasivan
Chennai Mathematical Institute.
03-07-19

---

Abstract

Many recent innovations have led to software services that generate vast amounts of data. Leveraging this data is accomplished through machine learning. Quite often, we need to resort to sophisticated machine learning techniques to create applications that make a desired impact. Sophisticated machine learning techniques are computationally expensive. The size of data over which machine learning is performed makes naive machine learning approaches intractable. Application of machine learning to big datasets is an active area of research. This is also the focus of this dissertation. In this dissertation, three approaches to performing machine learning over big datasets are presented. These examined the application of non-parametric machine learning methods to big datasets. Each of the approaches reported in this thesis has attractive features. The appropriate choice should be guided by the characteristics of the problem and the desired model characteristics.