How to get started with Machine Learning - For Developers

Gave this presentation on the Axel Springer Developer Day 2020 where I presented some tips to get started with Machine Learning for Developers.

Transcript

1. How to get started with Machine Learning – For Developers Dat Tran – Head of AI @ Axel Springer AI Axel Springer Developer Day 2020 29 april 2020 ~ berlin ~ @datitran
2. echo $(whoami)
3. echo $(whoami) Face2Face Hydroplaning Prediction Image Quality Assessment
4. echo $(whoami)
5. echo $(whoami)
6. Why this talk?
7. A lot of developers ask me how they can get started with Machine Learning (ML) on a ”regular” basis
8. Why do they do this?
10. https://www.springboard.com/blog/machine-learning-engineer-salary-guide/
12. https://www.tesla.com/autopilotAI
13. https://github.com/as-ideas/ForwardTacotron
14. https://www.deepl.com/translator
16. Smart phone Smart car Smart tv Everything is/will be “smart” & “connected” in the future
17. Question: How to get started?
18. Many ways to get started oBootcamps oFree online courses oBlog posts oBooks oFriends oetc…
19. Programming: o Python o PyTorch/TensorFlow o Numpy o Flask, Falcon o Airflow o SQL o Etc… Mathematics: o Linear Algebra (Vector, Scalar, Matrix Multiplication) o Calculus (Derivatives, gradient descent) o Probability theory Data: o Structured o Unstructured Domain: o Marketing o E-Commerce o Automotive
20. Online Courses
21. Online Courses https://www.coursera.org/learn/machine-learning o 11 weeks program with good introduction o Covers basics from definition of terms to some more advanced concepts o “Math” heavy with quizzes and implementations o Octave (similar to Matlab) o Old
22. Online Courses https://www.coursera.org/specializations/deep-learning o 5 courses around deep learning (Feed-forward, convolutional neural networks, recurrent neural networks) o Python ftw o Not “free”
23. Online Courses https://course.fast.ai/ o Learn ”end-to-end” ML project with a lot examples o Covers theory as well o PyTorch o No certificate but “who cares”
24. And many more…
25. Books
26. o Pattern Recognition and Machine Learning – Christopher Bishop o Deep Learning – Ian Goodfellow, Yoshua Bengio, Aaron Courville o The Hundred-Page Machine Learning Book – Andriy Burkov o Hands-on Machine Learning with Scikit-learn, Keras and TensorFlow: Concepts, Tools, and Techniques to Build Intelligent Systems o And many more…
27. Springer has released 65 Machine Learning and Data books for free (end of July): https://towardsdatascience.com/springer-has-released-65- machine-learning-and-data-books-for-free-961f8181f189
28. Practice Practice Practice
29. o Data science & ML competition o Provide a lot of interesting datasets to learn from o Notebooks are also good to learn from o Educational training competition: Titanic, house prices https://www.kaggle.com/
30. o ML meta- search with code o Lots of interesting papers https://paperswithcode.com/
31. Stay-up-to-date oTowards Data Science oReddit ML oTwitter/LinkedIn: follow people like hardmaru, OpenAI, Google AI, Axel Springer AI and me J oAnd many more…
33. Some tips o Learn the basics (math, statistics & ML theory) well o Start somewhere whether it’s Kaggle, implementing a research paper etc. o Learn how to do “end-to-end” ML and then get into the details o Focus on a specific topic first e.g. Deep Learning, Computer Vision or NLP and then expand (T-Shape) o AI is changing fast, take 30minutes everyday to read on new stuff o Practice, Practice, Practice