How to Organize an ML Study Jam - WTM WiDS 2019

I gave a talk on how to organize an ML Study Jam at the joint event of Women Techmakers & Women in Data Science Lightning Talks meetup.

There are tons of great resources, for example, courses from tensorflow.org, Udacity, Coursera, MIT, Fast.ai, and Kaggle Learn.

Hopefully this helps you whether you are organizing an ML Study Jam for your meetup, a study group or study ML on your own.

Transcript

1. How to Organize an ML Study Jam Best practices & learning resources Margaret Maynard-Reid ML GDE @margaretmz Women Techmakers & Women in Data Science Lightning Talks, 8/26/2019 1
2. “AI is the new electricity” - Dr. Andrew Ng, 4/28/2017 2
3. @margaretmz | @WTMSeattle | @GDGSeattle Why run an ML Study Jam? Because ML skills are in high demand ML is the “Software 2.0” (Andrej Karpathy), every engineer and non engineer should learn the basics of ML. 3
4. @margaretmz | @WTMSeattle | @GDGSeattle Before • Remind attendees of prerequisites prior to the event • Identify content Study Jams best practices During • Talks • Hands-on • Coaches • Demos • Swags After • Social media & photos • Follow-up 4 Single day or multiple day sessions My blog post on How to Organize a Study Jam
5. @margaretmz | @WTMSeattle | @GDGSeattle ML prerequisites Python & Python libs • Python basics • Numpy - low level math operations • Matplotlib - data visualization • Pandas - data manipulation Calculus • Derivative • Gradient • Chain rule Algebra & linear algebra Stats 5
6. @margaretmz | @WTMSeattle | @GDGSeattle Google Colab What is Google Colab? • Jupyter Notebook ◦ stored on Google Drive ◦ running on Google’s VM in the cloud • Free GPU and TPU! • TensorFlow is already installed • Save and share from your Drive • Save directly to GitHub Check out these learning resources • My blog on Colab • TF team’s blog on Colab • Laurence’ Video Build a deep neural network in 4 mins with TensorFlow in Colab • Paige’s video How to take advantage of GPUs & TPUs for your ML project • Sam’s blog Keras on TPUs in Colab 6
7. @margaretmz | @WTMSeattle | @GDGSeattle Study Jam content Where to find the ML content? From challenging to easy, mix and match these strategies Use existing content Write your own Invite an expert 7
8. @margaretmz | @WTMSeattle | @GDGSeattle Write your own content Pros: • Re-use your own code, tutorial and blog posts • Greatest flexibility Cons: • Time consuming 8
9. @margaretmz | @WTMSeattle | @GDGSeattle Blog.tensorflow.org TensorFlow on Youtube TensorFlow on Twitter #AskTensorFlow #TensorFlowMeets TensorFlow Dev Summit 2019 9 Start with tensorflow.org
10. @margaretmz | @WTMSeattle | @GDGSeattle MOOCs Udacity, Coursera and MIT all have deep learning course with TensorFlow 2.0 Fast.ai uses PyTorch and has two lessons on Swift for TensorFlow 2 month, intermediate 1 month, 4 courses 5 session. 9 lectures 2 parts 14 lessons link link link link 10
11. @margaretmz | @WTMSeattle | @GDGSeattle Machine Learning Crash Course Developed by Google Pros: • Great ML fundamentals • Nice animated explanations, for example - ◦ Optimizing learning rate ◦ Backpropagation • Complete ML glossary Cons: • TensorFlow code in outdated version 11
12. @margaretmz | @WTMSeattle | @GDGSeattle Qwiklabs 12 Go to qwiklabs.com and search for “ML”: • Baseline: Data, ML, AI • Advanced ML: ML Infrastructure • Intro to ML: Image Processing • Intro to ML: Language Processing • Intermediate ML: TensorFlow on GCP
13. @margaretmz | @WTMSeattle | @GDGSeattle Kaggle Learn 13
14. @margaretmz | @WTMSeattle | @GDGSeattle Idiomatic Programmer Andrew Ferlitsch https://github.com/GoogleCloudPlatform/keras-idiomatic-programmer 14
15. @margaretmz | @WTMSeattle | @GDGSeattle Recommended books By Aurélien Géron Deep learning with Python by Francois Chollet 15
16. @margaretmz | @WTMSeattle | @GDGSeattle How to read research papers • Arxiv.org • arxiv-sanity.com • Papers with code 16
17. Thank you! Follow me on Twitter, Medium & GitHub to learn more about on-device ML, TensorFlow & computer vision. @margaretmz @margaretmz margaretmz Margaret Maynard-Reid ML GDE @margaretmz North America GDE Summit 17