The Beauty that is the Delta Rule In general, there are 2 main ways to train an Artificial Neural Network (ANN). In our previous post , I have told you about the popular perceptron rule that has been around for a long time. We also said that the perceptron training rule is guaranteed to converge …
The Perceptron Training Rule It is important to learn the training process of huge neural networks. However, we need to simplify this by first understanding how a simple perceptron is trained, and how its weights are updated! Only then, will we be able to understand the dynamics of complicated and monstrous neural networks, such as …
ECML-PKDD-2019 on EBFDD networks for Anomaly Detection This paper introduces the Elliptical Basis Function Data Descriptor (EBFDD) network, a one-class classification approach to anomaly detection based on Radial Basis Function (RBF) neural networks. The EBFDD network uses elliptical basis functions, which allows it to learn sophisticated decision boundaries while retaining the advantages of a shallow …
ECML-PKDD-2019: Elliptical Basis Function Data Descriptor (EBFDD) for Anomaly Detection Read More »
What will you Learn in this Post? Neural Networks are function approximators. For example, in a supervised learning setting, given loads of inputs and loads of outputs in our training data, a neural network is capable of finding the hidden mapping between the those inputs and outputs. It can also, hopefully, generalize its acquired knowledge …
Train a Perceptron to Learn the AND Gate from Scratch in Python Read More »
Hello friends and welcome to MLDawn! So what sometimes concerns me is the fact that magnificent packages such as PyTorch (indeed amazing), TensorFlow (indeed a nightmare), and Keras (indeed great), have prevented machine learning enthusiasts to really learn the science behind how the machine learns. To be more accurate, there is nothing wrong with having …
This talk was delivered in PyData meetup. This was an amazing meetup with great talks and ideas being discussed! Thanks to organizers, I was also invited to give a talk. Below is the summary: In this talk we learn about what Artificial Neural Networks (ANNs) are, and find out how in general, Maximum Likelihood …
Birth of Error Functions in Artificial Neural Networks Read More »
Hello friends and welcome to MLDawn! So what sometimes concerns me is the fact that magnificent packages such as PyTorch (indeed amazing), TensorFlow (indeed a nightmare), and Keras (indeed great), have prevented machine learning enthusiasts to really learn the science behind how the machine learns. To be more accurate, there is nothing wrong with having …
What is this post about? In the previous post, we learned all about what over-fitting is, and why over-fitting to our training data could easily hinder the generalization power of the final trained model to the future unseen test data. In particular, we said that there are 2 cases where over-fitting can harm us: If …
What is this post about, and what is next? Hi there!!! Let’s cut to the chase, shall we?! You have this decision tree algorithm, coded beautifully through Scikitlearn (I applaud you!), or from scratch using numpy (I am proud of you!). You let the bloody thing train, and train and you see the training error …
Which attribute to choose? (Part-2) Today we are going to touch on something quite exciting. In our previous post we talked about Entropy of a set, E(S), and told you that entropy is nothing but a quantitative measure of how mixed up our set is! I also showed you that regardless of how the decision …