はてなブックマーク

In every business, legal contracts are foundational documents that define the relationships, obligations, and responsibilities between parties. Whether it’s a partnership agreement, an NDA, or a supplier contract, these documents often contain critical information that drives decision-making, risk management, and compliance. However, navigating and extracting insights from these contracts can be a

"Bigger is always better" – this principle is deeply rooted in the AI world. Every month, larger models are created, with more and more parameters. Companies are even building $10 billion AI data centers for them. But is it the only direction to go? At NeurIPS 2024, Ilya Sutskever, one of OpenAI’s co-founders, shared an idea: "Pre-training as we know it will unquestionably end". It seems the era o

What is the Problem with Data Engineering Today?If you search for a clear definition of what data engineering actually is, you’ll get so many different proposals that it leaves you with more questions than answers. But as I want to explain what needs to be redefined, I’ll better use one of the more popular definitions that clearly represents the current state and mess we all face: Data engineering

We’ve witnessed remarkable strides in AI image generation. But what happens when we add the dimension of time? Videos are moving images, after all. Text-to-video generation is a complex task that requires AI to understand not just what things look like, but how they move and interact over time. It is an order of magnitude more complex than text-to-image. To produce a coherent video, a neural netwo

Note: When I initially published this article back in February it was an early thought experiment. Since then I have started working on a book for "Generative AI Design Patterns" with a major publisher. Please follow me to keep updated on updates to my patterns an ideas in this space. The Need For AI Patterns We all anchor to some tried and tested methods, approaches and patterns when building som

Image created by author using Dall-E 3 Figure 1: Root Cause Workflows for LLM RAG Applications (flowchart created by author) If you have been experimenting with large language models (LLMs) for search and retrieval tasks, you have likely come across retrieval augmented generation (RAG) as a technique to add relevant contextual information to LLM generated responses. By connecting an LLM to private

The following article explains various methods for computing distances and showing their instances in our daily lives. Additionally, it… Various ML metrics. Inspired by Maarten Grootendorst. "There is no Royal Road to Geometry." – Euclid Quick note: Everything written and visualized has been created by the author unless it was specified. Illustrations and equations were generated using tools like

How to improve the performance of your Retrieval-Augmented Generation (RAG) pipeline with these "hyperparameters" and tuning strategies Tuning Strategies for Retrieval-Augmented Generation Applications Data Science is an experimental science. It starts with the "No Free Lunch Theorem," which states that there is no one-size-fits-all algorithm that works best for every problem. And it results in da

Imagine studying a module at university for a semester. At the end, after an intensive learning phase, you take an exam – and you can recall the most important concepts without looking them up. Now imagine the second situation: You are asked a question about a new topic. You don’t know the answer straight away, so you pick up a book or browse a wiki to find the right information for the answer. Th

RAG vs Finetuning – Which Is the Best Tool to Boost Your LLM Application? Image by author Prologue As the wave of interest in Large Language Models (LLMs) surges, many developers and organisations are busy building applications harnessing their power. However, when the pre-trained LLMs out of the box don’t perform as expected or hoped, the question on how to improve the performance of the LLM appl

Unlock advanced customer segmentation techniques using LLMs, and improve your clustering models with advanced techniques Content Table · Intro · Data · Method 1: Kmeans · Method 2: K-Prototype · Method 3: LLM + Kmeans · Conclusion Intro A customer segmentation project can be approached in multiple ways. In this article I will teach you advanced techniques, not only to define the clusters, but to a

The Quick-start Guide Isn’t Enough "Retrieval augmented generation is the process of supplementing a user’s input to a large language model (LLM) like ChatGPT with additional information that you (the system) have retrieved from somewhere else. The LLM can then use that information to augment the response that it generates." – Cory Zue LLMs are an amazing invention, prone to one key issue. They ma

Image created by the authors. How can we test applications built with LLMs? In this post we look at the concept of testing applications (or prompts) built with language models, in order to better understand their capabilities and limitations. We focus entirely on testing in this article, but if you are interested in tips for writing better prompts, check out our Art of Prompt Design series (ongoin

Overcoming token size limitations, custom model loading, LoRa support, textual inversion support, and more Stable Diffusion WebUI from AUTOMATIC1111 has proven to be a powerful tool for generating high-quality images using the Diffusion model. However, while the WebUI is easy to use, data scientists, machine learning engineers, and researchers often require more control over the image generation p

GPT-4 vs. ChatGPT: An Exploration of Training, Performance, Capabilities, and Limitations Image created by the author. OpenAI stunned the world when it dropped ChatGPT in late 2022. The new generative language model is expected to totally transform entire industries, including media, education, law, and tech. In short, ChatGPT threatens to disrupt just about everything. And even before we had time

(Image from Unsplash) This is the second article of our GPT series, where we will dive into the development of GPT-2 and GPT-3, with model size increased from 117M to a staggering 175B. In case you are interested in the other articles in this GPT series, check the links below: Part 1: Understanding the Evolution of ChatGPT: Part 1 – An In-Depth Look at GPT-1 and What Inspired It. Part 3: Insights

It appears that every sophisticated ML team has built a feature store for their ML platform. Uber built Palette. Airbnb built Zipline. Netflix built Time Travel. Google Cloud worked with our customer GoJek to build Feast. Fortunately, you no longer need to build or manage your own. Google Cloud Vertex AI offers a fully managed feature store as does Sagemaker. There are even companies like tecton.a

[Image by Author]One of the things I hear more often from data scientists is: “We have a problem: this dataset is imbalanced.” After telling this, they usually start proposing some balancing techniques to “fix” the dataset, such as undersampling, oversampling, SMOTE, and whatnot. I think this is one of the most widespread misconceptions in the machine learning community. So, I would like to state

Photo by Kai Dahms on UnsplashI love creating software libraries. Two months ago, I started porting one of our Python packages into a Rust crate. This new Rust crate matches the Python package’s ease of use and expressiveness. Along the way, I learned nine rules that can help you create beautiful libraries in Rust. The rules are: Create examples that don’t embarrass you.Accept all kinds of strings

Tuning deep learning pipelines is like finding the right gear combination (Image by Tim Mossholder on Unsplash) Why should you read this post? The training/inference processes of deep learning models are involved lots of steps. The faster each experiment iteration is, the more we can optimize the whole model prediction performance given limited time and resources. I collected and organized several

On 20th May 2021 Google held its developer conference I/O and announced a new algorithm for their search engine: MUM, a Multitask Unified Model [1]. For the last two years, BERT was the underlying model for their search engine. BERT was a breathtaking release and was state-of-the-art until now, until MUM came. The algorithm BERT changed a lot in the field of NLP and was applied in thousands or eve

Airflow is by default very permissive and without strict rules you are likely to create a chaotic code base that is impossible to scale and administrate. if you are not careful your shortcuts will cost you a lot afterwards Airflow permissive approach will let you schedule any custom code (jobs) but you will create a spaghetti stack if you do not follow very strict SEPARATION OF CONCERN design betw

Image by author. Table of contents Introduction Automated Exploratory Data Analysis packages 2.1 DataExplorer 2.2 GGally 2.3 SmartEDA 2.4 tableone Conclusions References 1. Introduction Exploratory Data Analysis (EDA) aims at performing an initial investigation on the data by summarizing their characteristics through statistical and visualization techniques, and it is a critical early step in any

Thoughts and Theory, State of the Art Digest It’s been quite a year for Graph ML – thousands of papers, numerous conferences and workshops… How do we catch up with so many cool things happening around? Well, we are puzzled as well and decided to present a structured look at Graph ML highlighting 🔥 trends and major advancements. The image was generated by ruDALL-E with a prompt "graphs floating in

(Image by Author) PyCaret’s New Time Series Module🚪 IntroductionPyCaret is an open-source, low-code machine learning library in Python that automates machine learning workflows. It is an end-to-end machine learning and model management tool that speeds up the experiment cycle exponentially and makes you more productive. In comparison with the other open-source machine learning libraries, PyCaret

Summary of CLIP model’s approach, from Learning Transferable Visual Models From Natural Language Supervision paper Introduction It was in January of 2021 that OpenAI announced two new models: DALL-E and CLIP, both multi-modality models connecting texts and images in some way. In this article we are going to implement CLIP model from scratch in PyTorch. OpenAI has open-sourced some of the code rela