A Logician’s Perspective Of Paradigms In Machine Learning

The BMI session on 14th October, was adorned by Dr. Ramyaa, Faculty of New Mexico Institute of Technology on “A logician’s perspective of paradigms in Machine Learning”. Our speaker is a logician who loves math, quizzes and loves solving puzzles and has done immense work in the field of AI, ML.

Our guest explained that the Paradigms are the underlying philosophy and it is changing very swiftly with the time and our favourite one might become unfavourable for practices. She unravelled the technical concepts for easy understanding and explained what Artificial Intelligence and Machine Learnings are and the differences between them. She explained that AI as any machine that simulates and mimics human traits of intelligence and smartness with rationality. AI is driven by data and it draws implicit and explicit rules from the data. Machine Learning is not what we teach a machine about something from something, but it is based on the assumptions of the underlying models, for example the gravitational pull witnessed by newton with the fall of Apple. The goal of ML is to learn such underlying models from data with gradient descent which is the heart of ML. She pointed out the the problems associated with ML as Wrong Abstractions, Catastrophic Disturbance, adverse attacks.

Dr. Ramyaa, expounded the AI winter caused by the challenges with the technology and the AI resurgence with the research and development in neural networks imitating to the neurons in the human brains. The future of AI is very versatile with the evolution of Interpretable AI, Differentiable logic learning. Our guest concluded by saying that the paradigms shift is inevitable but choose and focus on what you like and love to do so that we will be able to accommodate it within the transformations.