DFS can be implemented in two ways: iterative and recursive. Here, I’ll show you how to do it recursively as IMHO it is easier to understand and to code. This is also a fantastic opportunity to learn how recursion works if you’re not familiar with it yet. DFS implementation will be in pure Python. Below…

There is a lot of talk about the value of data being the new gold. Many companies are therefore pouring large sums of money into becoming data-driven. It is sold as a completely new way of doing business, almost as if business has never been driven by data before. Sure, many companies struggle with being…

Data Format Fundamentals — Single Precision (FP32) vs Half Precision (FP16) Now, let’s take a closer look at FP32 and FP16 formats. The FP32 and FP16 are IEEE formats that represent floating numbers using 32-bit binary storage and 16-bit binary storage. Both formats comprise three parts: a) a sign bit, b) exponent bits, and c)…

A gentle introduction to the latest multi-modal transfusion model Recently, Meta and Waymo released their latest paper — Transfusion: Predict the Next Token and Diffuse Images with One Multi-Modal Model, which integrates the popular transformer model with the diffusion model for multi-modal training and prediction purposes. Like Meta’s previous work, the Transfusion model is based on the…

One image can be worth thousands of words. Image by authorA confusion matrix is a convenient way to present the types of mistakes a machine learning mode makes. It is an N by N grid with numbers, where the value in the [n, m] cell represents the number of examples annotated with the n-th class…

Revisiting CPU for ML in an Era of GPU Scarcity Photo by Quino Al on UnsplashThe recent successes in AI are often attributed to the emergence and evolutions of the GPU. The GPU’s architecture, which typically includes thousands of multi-processors, high-speed memory, dedicated tensor cores, and more, is particularly well-suited to meet the intensive demands…

Enhancing LLM decision-making: integrating language agent tree search with GPT-4o for superior problem-solving Image by the author: midjourney — abstract puzzleLarge Language Models (LLMs) have demonstrated exceptional abilities in performing natural language tasks that involve complex reasoning. As a result, these models have evolved to function as agents capable of planning, strategising, and solving complex…

I am now bringing you a Data Science in Marketing — Customer Base Segmentation (clustering) project. This is a high-level project that can be applied in any company that needs to understand the differences among its customers. Know that companies are eager 👹 to understand who their customers are. Or do you think all customers…

Unlocking the power of large language models Photo by ZHENYU LUO on UnsplashIn this article, I will examine how large language models (LLMs) can convert natural language into SQL, making query writing more accessible to non-technical users. The discussion will include practical examples that showcase the ease of developing LLM-based solutions. We’ll also cover various…

User Persona based Personalization of LLM generated Responses ChatGPT, or the underlying large language models (LLMs) today, are able to generate contextualized responses given a prompt. As a next step in the LLM evolution, we expect the responses to be more and more personalized with respect to the persona, conversation history, current conversation context and…