Understanding the underlying algorithms and theories of machine learning

Abstract

Machine learning is largely known as the basis for Artificial Intelligence (AI). The name itself describes exactly that, machines that learn. We are seeing an increasing amount of these machines in our day-to-day lives as well but we may not even realise it. The machines involved in machine learning are not exactly the robots learning to walk or work assembly lines. In fact, machine learning is the study of algorithms and statistical models that computer systems use so that they can perform tasks without being told exactly what to do. The algorithms are designed to find patterns and inferences in the data generated by the task at hand and use those to not only complete the task but learn how to perform said task more efficiently. In the age of big data machine learning has become an essential tool in interpreting massive datasets and developing new ways to use them. From the recommendations music and video streaming services make, to self-driving cars and speech recognition software, anywhere huge amounts of data require analysis and mining, machine learning algorithms are humming away in the background. While this field has become essential to our modern way of life and transformative for many industries, it is also a new and rapidly evolving field. Designing new algorithms that can handle increasing amounts of data with greater complexity is one area of active research, as is refining the statistical models involved in machine learning. Professor Yoshinobu Kawahara, who is based at the Institute of Mathematics for Industry at Kyushu University in Japan is carrying out research dedicated to studying the underlying algorithms and theories of machine learning.