Python &
Notebooks
for AI
Why Python
• Python has become the most popular
programming language today by some metrics
• It’s a great higher-level programming language
– Easy to learn and use
– Many interesting and powerful features
– Large library of modules; easy to install
• Drawbacks: slow (interpreted); few built-in
datatypes (no arrays!), poor multiprocessing,…
• Overcome by new modules implementing
efficient data structures in C (e.g., Numpy)
Why Python for AI?
• AI is currently undergoing a Cambrian explosion
• New ideas appear every week and must be
evaluated and explored rapidly to maintain an
advantage or even keep up
– Applies to companies, researchers, students
• Python is great for rapid development
• New neural network packages (e.g., TensorFlow,
PyTorch) use efficient C modules for expensive
computing parts, visualization, etc.
Jupyter Notebooks
• Current Python notebook
software, replacing iPython
• Open source, browser-
based application to create and share
interactive documents with
– Live Python code (also R, Julia, Scala, Bash, …)
– Visualizations
– Narrative Text
• Also has a console window and file mover
Three ways to use…
• Install on your computer
– Pip install jupyter; cd to working directory; execute
command “jupyter notebook”; visit browser
• In your browser, access UMBC’s remote Jupyter
server, [Link]
– Requires authorization and then logging in
• Access Google’s free Colab server at
[Link]
– Powerful but harder to work with local files on your
computer
• Ok, there are more ways, but…
Let’s try it
Example: code to solve
water jug problems
with two jugs