Talks - Machine Learning/Deep Learning

Agenda:

1. Understanding Principal Component Analysis - Joydeep - 10:15 - 11:15

Principal Component Analysis is an important statistical procedure that we can undertake in machine learning and thus reduce the number of dimensions in our dataset. This helps in increasing the performance of our model, as well as in visualizing the dataset and is a great tool in case we suspect that the various features in the dataset are correlated to each other. In this session we will get to know the underlying mathematics of why PCA is a useful tool, when using it provides great benefit to us, and how to implement this in code.

Prerequisites:

a. Some statistical knowledge. How datasets are generated. What is a matrix? matrix multiplication and basic matrix operations and how they function.
related readings:

1. https://www.khanacademy.org/math/precalculus/precalc-matrices/intro-to-matrices/v/introduction-to-the-matrix

2. https://www.khanacademy.org/math/precalculus/precalc-matrices/multiplying-matrices-by-matrices/v/matrix-multiplication-intro

3. https://www.georgebrown.ca/uploadedFiles/TLC/_documents/Basic%20Matrix%20Operations.pdf

b. What is variance and covariance. related readings:

1. https://www.youtube.com/watch?v=G16c2ZODcg8 2. http://stattrek.com/matrix-algebra/covariance-matrix.aspx

c. python and how to run jupyter notebook. Please install anaconda3-4.4.0.
http://jupyter-notebook-beginner-guide.readthedocs.io/en/latest/execute.html

11:15 - 11:30: Break

Building your own neural network using nothing but python - Niranjan (11:30 - 12:30)

The best way to understand a neural network, is to build one from scratch without using TensorFlow or Scikit learn etc.. This talk will introduce you to all the math you will need to build a toy neural network of your own. While the neural network you build will not help you predict or classify with great accuracy, it does give you the solid foundation you need to explore this topic further.

Notes:

1. RSVP opens 7 days before the event.

2. The event is free of cost.

3. Waitlisted participants will receive confirmation notification about a day before the event.

4. If you aren't sure or have other important work to do, please UNRSVP and help others attend.

Follow us on Twitter (http://twitter.com/bangpypers) and FaceBook (https://www.facebook.com/bangpypers/). Join our discussion in Slack (http://bangpypers.heroku.com/) and Mailing List (https://mail.python.org/mailman/listinfo/bangpypers). Read about past meetup in the blog (http://bangalore.python.org.in/)