What Are You Up To?
Our Community Update for November
Hey Outliers,
We skipped this month’s meeting since it landed on Thanksgiving holiday in the US. So this month’s newsletter will focus on updates within our community!
COMMUNITY UPDATES
BOOK STUDY
Advanced Topics Group
Book: Bertsekas, Dimitri, and John Tsitsiklis. Introduction to Probability. 2nd ed.
Tackling the exercises together in Discord
Interpretable Machine Learning
On Section 8.4 Feature Decomposition
Working on the Python code with the goal of contributing to its GitHub repository
https://github.com/tqrahman/interpretable-ml-book/tree/master/Python
RESEARCH PAPERS
DarthVader had some great sessions going over Transformers. He took us back to the basics with Neural Machine Translation (the precursor to Transformers) and then went over Attention is All You Need! He provided his notes on the paper below.
Neural Machine Translation by Jointly Learning to Align and Translate (Notes)
Attention is All You Need (Notes)
KAGGLE
Melling reached Kaggle Notebooks Master! He is also working on Dataset Masters. Check out his dataset in Kaggle:
https://www.kaggle.com/datasets/mmellinger66/new-jersey-wineries
BLOG
How Should I Feed My Gradient Descent
We always heard that Mini-Batch is the way to go when calculating Gradient Descent. This blog shows different iterations of Gradient Descent and why ML practitioners lean towards Mini-Batch Gradient Descent.
We have a Medium Publication, Biased Outliers (formerly KaggleNYC). If you would like to submit your blog to the publication, let us know!
PROJECTS
StockGrab
A Streamlit app that web scrapes stock data based on user input to allow users to download the data. It also allows the user to download the data.
Coming Soon: Price Prediction using Time Series. If you would like to contribute to this project, let us know!
Operation Blossom
A Streamlit app that allows users to upload an image and identify the type of flower it is. In addition, it will show what model it used and its accuracy.
Project Roadmap: https://docs.google.com/document/d/1ToFOEp-HckYZaGHY3duvez7o8rmZOy2hx-7HnPtNWgk/edit
Coming Soon: An Interactive Streamlit App
EVENTS
DeepLearningAI Shows us how to build an AI portfolio
Boston Python has a friendly office hour to help answer Python questions
COURSES
Amazon is providing free AWS training here
Let us know if you would like to participate in any of our data science ventures! Most of our sessions happen in our Discord. We will not have a meeting on December due to the holidays, but we will keep the newsletter going. Until next time!
Featured Photo by Jed Villejo on Unsplash