One of my main motivations for writing this article stems from a particular client’s refusal to adopt the usage of Docker due to concerns related to Docker security and potential user permission escalation within the container itself. As a relatively naïve and green AI engineer doing ML deployment with Docker, I found myself unequipped with the knowledge to address this issue.
What is Docker and why is it so popular?
Since its release in 2013, Docker has gained massive traction amongst software companies as it made deployment of containerized microservices extremely convenient and easy. …
As an AI/ML Engineer (this applies to Software Engineers as well), the faster one can crunch out code, the greater our efficiency. Engineers are often good at their craft, but few try to squeeze out efficiencies in their day-to-day to maximize those gains.
These efficiencies or life hacks as I’d like to call them, may seem so insignificant to some but over time, really do add up to a lot. …
So, I decided to build a web application to classify durians because hey, why not? Check it out here.
To all my international readers, if you don’t know what a durian is, it is a fruit (in Singapore we call it the King of Fruits) that has a creamy texture, a pungent smell (see image below) and a spiky exterior. This said pungent smell makes people either hate it or absolutely LOVE it (I fall into the latter category, obviously). If you think it smells good, then it probably tastes even better.
Note: This article is largely inspired by Lee Kai-Fu’s ‘AI Superpowers’, along with some of my personal thoughts. Many ideas are abstracted from his book and any similarities found are intentional.
As a student, practitioner and advocate of AI, I’m fortunate to get a glimpse into how the AI technology works beneath the surface, and to a certain extent also allows me to imagine what kind of possibilities we have with AI not only in the present but also in the future.
To the man on the street, AI is seen as a technology that has the possibility to replace…
In recommender systems, we typically work with very sparse matrices as the item universe is very large while a single user typically interacts with a very small subset of the item universe. Take YouTube for example — a user typically watches hundreds if not thousands of videos, compared to the millions of videos YouTube has in its corpus, resulting in sparsity of >99%.
This means that when we represent the users (as rows) and items (as columns) in a matrix, the result is an extremely sparse matrix consisting of many zero values (see below).
In the context of recommender systems, Field Aware Factorization Machines (FFM) are particularly useful because they are able to handle large, sparse datasets with many categorical features.
To understand how FFM came about, let’s nail down some basics and understand why FFM are good and what they’re good for.
The simplest model we can think of when we try to model the relationship between a dependent variable and one or more independent variables is a linear regression model.
For example, to predict what ratings a user might give a particular movie, we could use many different features as predictors. …
Netflix is synonymous to most people in this day and age as the go-to streaming service for movies and tv shows. What most people do not know, however, is that Netflix started out in the late 1990s with a subscription-based model, posting DVDs to people’s homes in the US.
In 2000, Netflix introduced personalised movie recommendations and in 2006, launched Netflix Prize, a machine learning and data mining competition with a $1 million dollar prize money. Back then, Netflix used Cinematch, its proprietary recommender system which had a root mean squared error (RMSE) of 0.9525 and challenged people to beat…
This is the first of a three-part series to clustering, where I will cover some of the most popular clustering algorithms including K-Means, Agglomerative clustering and Gaussian Mixture Models. These are different clustering methods based on partition/distance, hierarchy and density respectively.
This article specifically covers K-Means clustering.
“The unsupervised learning is the way most people will learn in the future. You have this model of how the world works in your head and you’re refining it to predict what you think is going to happen in the future.” — Mark Zuckerberg
Unsupervised learning forms a very niche part of Machine…
In portfolio management, people often say that diversification is key to managing risk. While I do believe that (in some cases), for most retail (individual) investors, concentration in a handful of assets would make a lot more sense.
For most individuals, the amount of free cash (cash after expenses) they have on hand to invest at any point is low. This is often due to loans they have to pay off monthly — student loans, car loans, mortgage loans, etc.
As such, should an individual with $10,000 in free cash invest into 10 different stocks, with $1000 invested…
When I was an undergraduate studying economics, I didn’t know what I was going to do in the future. I had no notion of what my career would be like or what I had to do to get there.
All I knew was that I wanted to be successful.
Back then, success to me was defined as being sufficiently rich — to own a car, a house and other material goods that I wanted. It is to be in a state where money no longer matters. It is not about being flamboyant and flaunting one’s wealth, but to live a…