Friendships of AI: Discovering Hebbian Learning

Hello, dear readers! Today we delve into an intriguing concept in Artificial Intelligence (AI): Hebbian Learning. Borrowing directly from the way our brains function, Hebbian Learning promises to shape the future of AI. Hebbian Theory – The Networking Nature of Our Brains Our brain is a vast network of neurons, with each neuron being an individual in this network. Psychologist Donald Hebb proposed an idea about how this neural ‘social network’ functions. When neurons communicate frequently, their bond strengthens. Just like in human friendships, the more time spent together, the stronger the bond. Hebb summarized this principle as, “Neurons that […]

Deep Dive Into Capsule Networks: Shaping the Future of Deep Learning

In the realm of machine learning, traditional Convolutional Neural Networks (CNNs) have established a strong foothold, contributing significantly to image recognition and processing tasks. However, they’re not without their limitations, such as struggling to account for spatial hierarchies between simple and complex objects, and being heavily dependent on the orientation and size of the object. A newer framework, known as a “Capsule Network” (CapsNet), has been proposed to overcome these challenges. CapsNet, introduced by Geoffrey Hinton, Sara Sabour, and Nicholas Frosst in 2017, takes a different approach to object recognition and offers a promising alternative to CNNs. What are Capsule […]

Unraveling the Mystery of Evolutionary Neural Architecture Search: Simplification, Use Cases, and Overcoming Drawbacks

Introduction Evolutionary Neural Architecture Search (NAS) can be an enigma, even for those well-versed in machine learning and AI fields. Taking inspiration from the Darwinian model of evolution, evolutionary NAS represents a novel approach to optimize neural networks. This post aims to demystify evolutionary NAS, discuss its model mutations, delve into use cases, identify drawbacks, and provide alternatives. The Basics of Evolutionary NAS Just as biological species adapt and evolve over time, neural architectures can also ‘evolve’ to optimize their efficiency and effectiveness. Evolutionary NAS utilizes the principles of evolution—mutation, recombination, and selection—to automatically search for the best neural architecture […]