The Transformer architecture has revolutionized Artificial Neural Networks (ANN), demonstrating unparalleled performance in natural language processing, speech recognition, image processing, and other areas that require efficient handling of sequence data. This architecture is the main catalyst behind the current boom in generative AI, which made models like GPT and Stable Diffusion mainstream names. In this note we will take a deep look into the original implementation of the transformer model, and will then expand it to some of the modern variants.

Docker is just one of those technologies that make you fall in love 💕 with it, and make you wonder how you’ve manged to live your life without it for so long. It has revolutionized the way we develop, deploy, and run applications. On the surface, Docker is just a tool that allows us to create reproducible environments. But this simple concept is executed so well that it manages to completely change how you approach app development and service deployment as a whole. This note covers the fundamental concepts behind this wonderful tool and guides you through its basic usage.

Before we can process language-based data using Artificial Neural Networks (ANNs), we need to convert this data into some kind of a numerical representation. Embedding models are designed for this purpose. They transform language data into dense high-dimensional vectors that preserve the semantic associations between words. These vectors capture the essence of language data in a way that computers can understand and process. This note explores the most popular embedding model architectures, looks into how these models are trained, and discusses their critical role in Natural Language Processing (NLP).

If you are using Anaconda, the popular python distribution for data science, you know that it comes with its own package and environment manager called conda.

There are many tutorials out there, that tell you to update your conda environments using conda update --all. Luckily, you know that this is a terrible idea!

Decentralized clinical trials (DCTs) represent a transformative approach in the field of clinical research. By leveraging technology and innovative methodologies, DCTs offer a way to conduct studies that are less reliant on traditional, centralized site-based models. This shift is not just a logistical change; it’s a strategic one that can significantly impact drug development, patient accessibility, and the overall speed and efficiency of bringing new therapies to market.

Clinical trials are crucial in medical research, helping to determine the safety and efficacy of new treatments. The designs of these trials vary based on their objectives and the phase of the trial. In this note, we’ll explore some of the most popular and impactful clinical trial designs, from the classic randomized controlled trials to innovative adaptive designs.

Anyone involved in clinical trials would be familiar with the term “endpoint”, but it is surprising how many clinical professionals cannot clearly define what the term means or what differentiates an endpoint from an objective. This note is intended to bring some clarity on the matter.

Not all Artificial Neural Networks (ANN) are made the same, each is designed to tackle a specific type of problem. In this note, we’ll explore the most prominent architectures in detail, and learn how they are used in different applications.

Artificial Neural Networks (ANNs) exist in a fascinating intersection of neuroscience and machine learning, giving rise to powerful systems that can emulate intelligent behaviors. This note covers the essential glossary and theory behind this powerful technology.

Matrix factorization (or decomposition) is a fundamental concept in linear algebra that has widespread applications in various fields, including machine learning, data mining, and signal processing. At its core, matrix factorization involves decomposing a matrix into a product of two or more matrices, revealing the underlying structure of the data represented by the original matrix. This technique is especially powerful in uncovering latent features and reducing dimensionality, making it a cornerstone in many modern algorithms and systems.