Developing an understanding of a variety of LLM benchmarks & scores, including an intuition of when they may be of value for your purpose 17 min read · 14 hours ago Title card created by the authorIt seems that almost on a weekly basis, a new large language model (LLM) is launched…

Hidden dangers in seeking medical advice from LLMs Image by Author. 2 sub-images generated by DALLE-2Last year, ChatGPT passed the US Medical Licensing Exam and was reported to be “more empathetic” than real doctors. ChatGPT currently has around 180 million users; if a mere 10% of them have asked ChatGPT a medical question, that’s already…

Let’s create Python data objects in a few lines of code! Image generated by the author using Gencraft.If you’re a Python coder, you’re probably familiar with Zen. Three of its 19 guideline principles state that “explicit is better than implicit,” “readability counts,” and “simple is better than complex.” When you’re creating or integrating an existing…

Exploring the path to fast nearest neighbour search with Hierarchical Navigable Small Worlds 13 min read · 19 hours ago Image created by DALL·E 2 with the prompt “A bright abstract expressionist painting of a layered network of dots connected by lines.”Hierarchical Navigable Small World (HNSW) has become popular as one of…

In this short piece, I use public Wikipedia data, Python programming, and network analysis to extract and draw up a network of Oscar-winning actors and actresses. All images were created by the author. Wikipedia, as the largest free, crowdsourced online encyclopedia, serves as a tremendously rich data source on various public domains. Many of these…

Some of the struggles I face frequently as a data scientist Photo by ThisIsEngineering from Pexels: https://www.pexels.com/photo/female-software-engineer-coding-on-computer-3861972/Ostensibly, it may seem that being a data scientist is all sunshine and rainbows (at least I think that is the perception I give from my posts!). High pay, great benefits, flexible hours, and interesting work are some things…

An end-to-end implementation of a Pytorch Transformer, in which we will cover key concepts such as self-attention, encoders, decoders, and much more. 17 min read · 18 hours ago Photo by Susan Holt Simpson on UnsplashWhen I decided to dig deeper into Transformer architectures, I often felt frustrated when reading or watching…

Choosing the model that works best for your data We’ll use the EU AI act as the data corpus for our embedding model comparison. Image by Dall-E 3.OpenAI recently released their new generation of embedding models, called embedding v3, which they describe as their most performant embedding models, with higher multilingual performances. The models come…

Discover how FinalMLP transforms online recommendations: unlocking personalized experiences with cutting-edge AI research This post was co-authored with Rafael Guedes. The world has been evolving towards a digital era where everyone has nearly everything they want at a click of distance. These benefits of accessibility, comfort, and a large quantity of offers come with new…

Let’s begin with setting out what fine-tuning should do from a high level. Once you have a pre-trained a model to have strong generative capacities, you typically want to control its output somehow. Whether that be optimizing it to respond in dialogue as a chat-bot or to respond in code rather than English, the goal…