About this Event
The DMS Science Laboratory Committee and the Scientific Computing SIG once again presents a lecture series on deep learning. By the end of the series, we hope to have a solid base of Deep Neural Network power users at DMS that will serve as a basis for future scientific computing projects at the space.
This time we are going to try a different tactic. The last lecture series did a bottom-up approach where we started with some mathematical foundations. This time we are going to try the lecture series by starting with multilayer perceptrons, basic neural net architectures, and some pytorch workflows. We will then progress further into the details of these models as the lecture series progresses. You can think of this lecture series as doing the last lecture series in reverse order.
All who wish to attend are welcome!
Special Considerations and Warnings
First, I believe that you will get the most out of this session if you bring a laptop with anaconda and python installed. Download an installer from https://www.anaconda.com/products/individual
Here's an installation guide for PyTorch. If you have a decent NVIDIA graphics card, then select "CUDA 11.3" on the computing platform section of menu on the page, else select "CPU".
Check out this 60 minute course on PyTorch to help you get aquainted with the library
Just like last time,it is highly recommended that you read chapters 1 and 2 of https://d2l.ai/
If you want to learn linear algebra for later in the lecture series, then I still suggest that you check out the linear algebra section on the science laboratory committee page of the DMS wiki for additional supplemental materials https://dallasmakerspace.org/wiki/Category:Science#Linear_Algebra
Try to arrive on time as I plan on using the full 2 hours.