Mirroring the Human Brain: Artificial Neural Networks
An artificial neural network (ANN) is a network inspired by biological neural networks (as in the case of animals with a central nervous system and in particular, the brain) which are used to approximate functions that can depend on a large number of inputs, most of which are unknown. This is achieved by the thorough understanding of machine learning and cognitive science. All of this can easily confuse and intimidate even an engineering student. In order to make this now-widespread technology an area of interest for the students of MIT, the International Society of Automation (ISA) took upon themselves the task of conducting a workshop on the same. A three-day workshop, the third year students conducting it made sure that the entire teaching process was well-planned and coordinated.
The first day of the workshop, being held on a Sunday along with a bare minimum entry fee, attracted many students. The attendees were mostly first and second year students. Initially, a lot of participants, especially freshmen, seemed perplexed because of all the scientific terms that were being thrown around in the first quarter of the hour. But soon, things began to fall in place as each concept was explicitly explained with numerous examples from daily life applications. After every hour, a short break was taken in order to allow the new concepts to sink into the minds and to accommodate a one on one doubt-solving session.
The second and third day of the workshop saw a slight dip in the turnout, probably due to hectic schedules of full day classes and fest preparations. Nonetheless, the workshop was held in the most enthusiastic manner. These two days witnessed the elaboration of the topics touched upon on the first day. Mathematical proofs and real world applications of Multi-layer Perceptron, Auto-encoders, machine learning concepts like decision trees and random forests were covered.