Digests » 83
This document is based primarily on the presentation of feedforward networks in Pattern Recognition and Machine Learning by Christopher M. Bishop and an object-oriented design from David Selby who likewise was inspired by Denny Britz. In an algorithms course for my M.S. in Data Science, I was able to successfully program a variety of different algorithms but became frustrated with progress on the feedforward neural network. In the end, I only submitted a solution to the assignment utilizing a neural network package’s function.
Previously, when I was a student still studying at university, I wanted to get a data science internship. I wanted to get it so badly that I applied to so many companies, but only got a few replies from them.
In this post we'll explore what happens within a neural network when it makes a prediction. A neural network is a function that takes some input and produces an output according to some desired prediction. It's possible to make state-of-the-art predictions without understanding the concepts highlighted in this post. That's part of the beauty of modern computing and aggregate knowledge in general. But some things are too fundamental to just accept as fact and when I really stopped to look at these functions, machine learning became a little bit less of black box.
In this paper, we propose an elegant solution that is directly addressing the bottlenecks of the traditional deep learning approaches and offers a clearly explainable internal architecture that can outperform the existing methods, requires very little computational resources (no need for GPUs) and short training times (in the order of seconds). The proposed approach, xDNN is using prototypes.
A schism lies at the heart of the field of artificial intelligence. Since its inception, the field has been defined by an intellectual tug-of-war between two opposing philosophies: connectionism and symbolism. These two camps have deeply divergent visions as to how to "solve" intelligence, with differing research agendas and sometimes bitter relations.