Multi-label classification is a useful functionality of deep neural networks. I recently added this functionality into Keras' ImageDataGenerator in order to train on data that does not fit into memory. This blog post shows the functionality and runs over a complete example using the VOC2012 dataset.
We all remember the boom of Internet companies in the late 90s, then in the late 2000s mobile companies took center stage and have been dominating ever since. A new type is taken the spotlight, this is the era of AI companies, and like it has been before there are two options: adapt or fade away.
This week I was at the Spark+AI Summit 2018 conference in San Francisco. This post is a summary of my experience and highlights of the talks I attended.
In today's high pace user experience it is expected that new recommended items appear every time the user opens the application, but what to do if your recommendation system runs every hour or every day? I give a solution that you can plug & play without having to re-engineer your recommendation system.
Co-author of article written by Ronald van Loon about the impact of Deep Learning on the customer experience
Inspired by a new biological scientific research, I propose, build and train a Deep Neural Network using a novel neuron model.
With deep learning applications blossoming, it is important to understand what makes these models tick. Here I demonstrate, using simple and reproducible examples, how and why deep neural networks can be easily fooled. I also discuss potential solutions.
The need of the business to interact and understand the output from custom built machine learning models is increasing, here I provide an application skeleton to do just that with your Python made models.
Quick look into the Prophet API for predicting the number of transactions in a shop.