Human intelligence consolidates over continuous learning from experiences. The ability of the human brain to take an optimal decision in a complex situation is subjective to the individual intelligence level. Artificial intelligence, a global phenomenon, is harnessing the explosion of digital data and computational power with advanced computational algorithms. But where does Machine Intelligence fits into all of this?
Machine Intelligence and its application in IT world:
Machine Intelligence is a technology that integrates machine-learning algorithms with artificial intelligence. In the pursuit of machine intelligence, “Learning”, i.e. computer algorithms improving automatically with the experience, is central to the current global research on IT. In today’s era of Information Technology, there is an abundance of data being generated by different sources.
The development of computations based methodologies for learning data representation remains as the aim of scientists working worldwide both in academia and industry. As the modern IT systems are increasingly becoming reliable tools, a quick and accurate decision based on the mathematical analysis of big and dynamic data with the assistance of machine intelligence could make a huge difference to human life especially in health, water, security, manufacturing, and business sectors.
The data in any of the forms of text/ numeric/ signal/ image requires an objective evaluation to understand and potentially to modulate different aspects of human life. The data inherits uncertainties from sources and thus prompts the researchers to apply probability and fuzzy theories in machine learning as means to handle uncertainties. Many engineering, biology, economics, and social science applications involve uncertainties and imprecise data. The uncertainty in machine learning is mathematically represented in the form of randomness and/or fuzziness. While statistics and probability theory have been contributing in machine learning for decades, the applications of fuzzy concepts in machine learning is a research area whose potential still remains relatively unexplored. A principled handling of uncertainties (i.e. taking mathematically into account the uncertainties in a sensible manner) is one of the core issues of machine learning. It is observed that despite numerous publications of research results in the field of artificial intelligence and machine learning, some of the fundamental concerns regarding the handling of non-statistical-uncertainty while learning data representation remains unaddressed.
Novel fuzzy approach to quantification of uncertainties:
An analytical fuzzy theoretical approach to the machine learning has been recently introduced. The analytical quantification of the uncertainties on variables and nonlinear functions associated to the layers of a data-model by means of membership functions is the novelty of the approach. This research contributes to the field of fuzzy machine learning by a mathematical theory, nevertheless, offering a practical learning algorithm and providing a proof-of-concept through experiments. The results, built on the extensive research experience over the several years, includes the development of the mathematical framework of fuzzy theoretic machine intelligence algorithms with applications in machine learning, signal modeling & analysis, computer vision, process identification, and optimization. The undertaken research works have delivered intelligent computational solutions to address challenging problems in real-world. The practical implementations of research results have found applications in wide areas including Automation, mHealth & eHealth, Preventive Medicine & Public Health, Water System, Chemistry & Drug Design, and Environmental Sciences.