AI and Data and Stuff

Image generated using AI (Stable Diffusion)   Since I discovered the GPT-3-based chatGPT , I have become addicted to it and have so far resisted the urge to fill this blog with chatGPT posts but my resistance seems to wane with every successive interaction so I can't say how long I will be able to resist. I have interacted extensively with the model and, like many other people , have been completely blown away, especially when I remember what AI-generated content looked like few years ago when I used GPT-2 to generate a weird version of the 12 days of Christmas .  I have even attempted to hack or sidestep some of the baked in rules and content safeguards of chatGPT and so far my results have been a mix of successes and failures but my greatest win was when I finally got it to actually write code for various computer viruses and trojans (shhh don't tell anyone). This article, though, is not about hacking chatGPT but about another subject with also tickles my fancy: games. I wa...

Natural language processing (or NLP) is a branch of machine learning that deals with the languages spoken by humans. This branch includes things like automatic translation, text and speech generation amongst other things. The thing about most of the current NLP systems is that they require humongous amounts of sample text, speech and/or translations and other samples of the languages in focus to be able to be trained to perform these tasks of translation, generation etc This need for large sample data is the biggest challenge for many African languages because a majority of these languages have very limited amounts of such samples - too little to successfully train an NLP system. Many of these so called "low-resource" languages both in Africa and other parts of the world have been virtually overshadowed by the more popular languages like English, French, German, Spanish, Mandarin etc which have massive amounts of sample language data already available for use in trai...

Here is an exercise for you: Can you spot the fake face (or faces) among the six images below? By fake I mean generated (in other words, the person does not exist and has never existed). See the faces in larger format below (click on the image for an even higher resolution):              If your answer was one (or more) out of A, B, C, D, E or F then you are right. Actually ALL the faces are fake and the people in the images do not exist, have never existed and most likely never will. The faces were "imagined" and created (in high definition) by an Artificial Intelligence algorithm. Specifically it was created by a type of Deep Learning algorithm/model known as a GAN which is short for Generative Adversarial Network. What is a GAN? In concept, a GAN is a combination of two separate neural networks that compete against each other to ultimately produce "fake" data that is as close as possible to real sa...

Deep Learning is a branch of Machine learning and is something which I am hoping to master as soon as possible. I would have preferred to study it in a real university but due to lack of funds, I'm going the self-study route via online courses and books. When I first started on my Deep Learning self study journey, I started learning from the mathematical foundations because I read in many places that it was a good place to start and it made sense - deep learning at its foundation is really mathematics so why not start from the foundation and build upon that knowledge? I found a "mathematics for deep learning" curriculum online then went on youtube and started watching various videos on the topics in the curriculum - probability, statistics, calculus etc At first I really wanted to go deep with this foundation because I appreciated how important it was to thoroughly understand these foundations if I wanted to be an excellent machine learning engineer. Unfortunatel...

I recently joined a small project to produce an easily accessible COVID-19 dataset for Nigeria broken down into states. It all began while I was trying to find COVID-19 datasets for Nigeria broken down state by state, I realised that the only one available was in PDF format which makes it more challenging to extract just the data. I pitched the idea of a small project to extract the datasets from the PDF to CSV files to the nice folks at NG44 and they graciously agreed to support the project with funds. To be clear, we are not generating the data per se. That credit goes to the Nigeria Center for Disease Control (NCDC).  What we are mostly doing is manually extracting the data from the original PDF file format in which they are published into the more useful CSV file format and put it up on github. Github Link: Here's the shortlink to the dataset on github : https://bit.ly/covidnigeriadata About This dataset contains the official daily data for COVID-19 cases in Nig...

I have been brainstorming various ways of raising money for my Machine learning self-study courses and eventual Masters Degree and I recently came up with the idea of designing machine learning-themed T-Shirts and hoodies and selling them via Merch by Amazon on-demand printing service. I must confess I'm not a great designer and have never designed a T-shirt before but I have been able to come up with some simple designs. One of them you can see at the beginning of this article: "GANimal". Not sure how people will respond to my designs but I'm willing to take a shot at this and see where it goes. Here are more of my first designs, hope you like them (I'll post some more in the future):  So what do you think about them? Got any suggestions?

  As I learn machine learning, one thing I find fascinating (and also greatly enjoyable) is playing around with demos and pre-trained models of various machine learning algorithms and models. While many of the more recent models generally perform pretty well for many of the tasks they are built for, I am more interested in the relatively small area of the graph where the model goofs i.e messes up. Depending on the model and what it is supposed to do, these goofs can go very wrong, giving you results that range from weird, absurd and uncanny to hilarious and outright comical. From time to time I will post some of these goofy mistakes I experience while playing around with these demos and pre-trained models. Today, I present "The 12 days of Christmas" as generated by a GPT-2 model trained by Hugging Face using papers from Arxiv.org The 12 Days of Christmas is a popular Christmas song and Hugging Face is a company that specializes in creating AI and machine learni...