ISSUE 444 · July 18, 2023Posts & TutorialsGetting started with Code InterpreterOpenAI's Code Interpreter is a general-purpose toolbox that gives GPT-4 superpowers. With Code Interpreter, GPT-4 can ingest up to 100MB of your data and then use that data in python scripts that it writes and executes. That allows GPT-4 to do all sorts of things it couldn’t do before. This post is a nice tour of what's possible now and how to use it. VScode + Docker + Python= ❤️ ❤️ ❤️This is a great guide for setting up a Python development environment with VScode and Docker. It starts with a section that explains the advantages of each tool and how they work well together. From there, it's an easy to follow, step-by-step tutorial for setting everything up. Introduction to dimensionality reductionDimensionality reduction helps to simplify complex datasets and make them more manageable to work with. In this two-part introduction, Gabe Flomo shows how dimensionality reduction works and offers practical examples for common algorithms using python. Hex | Gabe Flomo Sponsored LinkComputational linguistics in the age of large language modelsWhat are the challenges facing LLMs? Amazon senior principal scientist and ACL 2023 general chair Yang Liu highlights the problem of “hallucination”, or generating false assertions, and explains how scientists are trying to address it. Tools & CodeAn Open-source Plotting Library for Statistical DataLets-Plot is a python plotting library for statistical data. It's based on the Grammer of Graphics and largely follows the gpplot2 API. This looks like a very nice plotting library that supports a wide range of chart types and features, such as data sampling, formatting, geocoding, notebook compatibility, and much more. Introduction to theft{theft} is an R package that provides a structured analytical workflow for the extraction, analysis, and visualisation of time-series features. It's designed to be flexible and extensible and it provides standardized access to key R packages as well as python libraries. PapersRegulating Frontier AI: To Open Source or Not?Two important new papers grapple with how to govern emerging and increasingly powerful "frontier AI" models. The first paper is a collaboration between Big Tech players like OpenAI, Google, Microsoft, etc. It argues for self-regulation with government oversight. The second paper, by Jeremy Howard, counters with an open-source approach. This is a great post that thoughtfully summarizes the issues. Financial Machine LearningThis new survey paper explores the nascent literature on machine learning in financial markets. Along with an extensive review, the paper highlights the best examples of what this line of research has to offer with recommendations for future research. ResourcesAdvanced Python MasteryThis course introduces Python's more advanced features and is intended to help you understand how to control the behavior of the language and bend it in ways that serve the needs of your application. This is an exercise-driven course that's free and self-paced. R for Data Science (2e)A new, second edition of R for Data Science was recently released and it's a big update. In this edition, visualization is more thoroughly covered, the programming section has been rewritten to focus on function writing and iteration, and there's a new section for accessing data from databases, spreadsheets, and the web. Free to read online. |