Editor's Picks

• Jürgen Schmidhuber's Annotated History of Modern AI and Deep Learning
  Machine learning is the science of credit assignment: finding patterns in observations that predict the consequences of actions and help to improve future performance. Credit assignment is also required for human understanding of how the world works, not only for individuals navigating daily life, but also for academic professionals like historians who interpret the present in light of past events. Here I focus on the history of modern artificial intelligence (AI) which is dominated by artificial neural networks (NNs) and deep learning, both conceptually closer to the old field of cybernetics than to what's been called AI since 1956...

• Cats, Pi, and Machine Learning [Project]
  There is a cat that wanders around my neighbourhood. I wanted to build something that would notify me whenever it came to my backyard...I thought to myself, if I can get a picture of my backyard as the input I can process the image by checking if there is a cat in it and send a notification to my phone as the output...For the picture input, I attached a camera module to a raspberry pi 4. For the processing, I wrote some code that would run on the pi, which would periodically take an image and run object detection on it. For the output, I relied on sending a message via Signal to my phone...

• Another Year of #TidyTuesday
  After another 52 data visualisations created for #TidyTuesday, it's time for the annual round-up! Read this blog post for some interesting R packages discovered, a few new I've tricks learnt, and the data visualisations I'd like to do again...

A Message from this week's Sponsor:

• Data Science on Kubernetes
• Hadoop replacement
• Apache Livy alternative
• Integration with Jupyter and Apache Zeppelin

Data Science Articles & Videos

• Annotated Forest Plots using ggplot2
  rstats blog post! a walkthrough on making forest plots and adding effect size annotations using ggplot2...

• The Build vs. Buy Guide for the Modern Data Stack
  Nishith Agarwal, Head of Data & ML Platforms at Lyra Health and creator of the popular open source data management framework, Apache Hudi, outlines his blueprint for building (and buying) the data stack of your dreams. In Part I of this series, Nishith discusses some initial considerations and shares a framework for getting started...

• Forward-Forward Algorithm App
  This app implements a complete open-source version of Geoffrey Hinton's Forward Forward Algorithm, an alternative approach to backpropagation...The Forward Forward algorithm is a method for training deep neural networks that replaces the backpropagation forward and backward passes with two forward passes, one with positive (i.e., real) data and the other with negative data that could be generated by the network itself...

• Large Language Models Encode Clinical Knowledge
  We present MultiMedQA, a benchmark combining six existing open question answering datasets spanning professional medical exams, research, and consumer queries; and HealthSearchQA, a new free-response dataset of medical questions searched online...we evaluate PaLM (a 540-billion parameter LLM) and its instruction-tuned variant, Flan-PaLM, on MultiMedQA. Using a combination of prompting strategies, Flan-PaLM achieves state-of-the-art accuracy on every MultiMedQA multiple-choice dataset (MedQA, MedMCQA, PubMedQA, MMLU clinical topics), including 67.6% accuracy on MedQA (US Medical License Exam questions), surpassing prior state-of-the-art by over 17%...

• Partially renaming columns using a lookup table
  Usually data sets come with short column names, which makes it easy to clean and manipulate the data. However, when presenting the data to stakeholders, in form of tables or plots, we often need longer, meaningful names. In many cases we have a lookup table which contains long and short versions of the column names so that we can “easily” replace the names when needed...Below we’ll look at how to rename columns using different approaches in R...

• ChatBCG: Generative AI for Slides [Twitter Thread]
  This Christmas Joseph Semrai and I finally got it working!! After DALL-E 2 for images and ChatGPT for text, the final step to make all of us redundant: The world’s first Text-to-PowerPoint AI...

• Top Python libraries of 2022
  Welcome to the 8th edition of our Top Python Libraries list!...We are excited to present this year's picks for the most innovative developments in the Python ecosystem. From this edition, we are expanding our list to include not only libraries per-se, but also tools that are built to belong in the Python ecosystem — some of which are not written in Python as you’ll see...

• Ask HN: Upskilling as a Data Engineer [HN Discussion]
  What should a data engineer learn as a part of Upskilling in 2022/2023?...New languages like Rust/Ocaml/Nim...if yes then which?...I don't think learning an ETL tool will be helpful because essentially they are all one and the same...Any tips?...

• Vanishing Gradients Podcast #15: Uncertainty, Risk, and Simulation in Data Science
  Hugo speaks with JD Long, agricultural economist, quant, and stochastic modeler, about decision making under uncertainty and how we can use our knowledge of risk, uncertainty, probabilistic thinking, causal inference, and more to help us use data science and machine learning to make better decisions in an uncertain world...This is part 1 of a two part conversation. In this, part 1, we discuss risk, uncertainty, probabilistic thinking, and simulation, all with a view towards improving decision making and we draw on examples from our personal lives, the pandemic, our jobs, the reinsurance space, and the corporate world. In part 2, we’ll get into the nitty gritty of decision making under uncertainty...

• University Professor Catches Student Cheating With ChatGPT
  A South Carolina college professor is raising the alarm after discovering that ChatGPT, a chatbot with artificial intelligence, was being used by a student to create an essay....

• At the forefront of AI with Hugo Larochelle, Research Scientist at Google
  On the show, we spoke about: His love of teaching and starting a YouTube channel. The rapid advances of the past 12 months. The illusion of AGI Enabling new scientific discoveries, Philanthropy, and the generous $1M donation from Hugo and his wife to combat climate change...

• What tech books (or courses?) have really helped you this past year? [Twitter Thread]
  It's the end of the year and some people have education budgets to spend! What tech books (or courses?) have really helped you this past year?...

Tool*

• built-in visualizations (like the image panel),
• custom project views, and
• your own python panels,

Tool*

Jobs

• Data Scientist / Machine Learning Engineer - Epsilon - NYC
  
  Epsilon Strategy and Insights, Data Sciences team is looking for a talented team player in a Data Scientist/Machine Learning Engineer role. You are an expert, mentor and advocate. You have strong machine learning and deep learning background and are passionate about transforming data into ml models. You welcome the challenge of data science and are proficient in Python, Spark MLLib, Tensorflow, Keras, ML algorithms and Deep Neural Networks, Big Data. You must be self-driven, take initiative and want to work in a dynamic, busy and innovative group...
   

Training & Resources

• Transformers from Scratch
  I procrastinated a deep dive into transformers for a few years. Finally the discomfort of not knowing what makes them tick grew too great for me. Here is that dive...Transformers were introduced in this 2017 paper as a tool for sequence transduction—converting one sequence of symbols to another. The most popular examples of this are translation, as in English to German. It has also been modified to perform sequence completion—given a starting prompt, carry on in the same vein and style. They have quickly become an indispensible tool for research and product development in natural language processing...

• Xavier Glorot Initialization in Neural Networks — Math Proof
  Detailed derivation for finding optimal initial distributions of weight matrices in deep learning layers with tanh activation function...

• An overview of gradient descent optimization algorithms
  Gradient descent is the preferred way to optimize neural networks and many other machine learning algorithms but is often used as a black box. This post explores how many of the most popular gradient-based optimization algorithms such as Momentum, Adagrad, and Adam actually work...

Last Week's Newsletter's 3 Most Clicked Links

• Why Business Data Science Irritates Me

• Everything I learned about accidentally running a successful tech conference

• Build a GPT-3 app: How I used GPT-3 to make gifting easier

Cutting Room Floor

• Creating in The Era of Creative Confidence
• Share what you do as a data engineer on a day to day basis? [Reddit Discussion]