Open spestana opened 4 months ago
Thank you for doing this, @spestana. I have a quick question: Do we assume everyone is familiar with ML basics and then move on to the methods (XGBoost, ANNs, etc)?
@Ibrahim-Ola Thank you, Ibrahim, for these questions. We are planning a series of one-on-one meetings with the tutorial leads to discuss specific details. I am generally available Monday, Thursday, and Friday mornings. Please let us know what works for you.
@NCristea, pardon me for responding late. Friday morning is a good time for me.
@Ibrahim-Ola—Great. Would this Friday, the 5th, work for you, or are you taking some time off?
@NCristea, this Friday works for me.
Meeting on 7/5/24 Ibrahim, Nicoleta, Ziheng, Jessica, Anthony
Q: Will we assume people are coming in with basic ML knowledge? A: Not necessarily, so it will be important to point people to introductory content in the week before the hackweek to help people prepare. We will not have time to teach the basics during the hackweek.
Our overall approach is to show what is possible with these tools and motivate people to explore. But to also say that if you want to use this for research, it's important to dig deeper and understand how things work, how to test and validate.
Lead: Ibrahim Alabi Date: 21/08/2024 Start Time: 1045 Duration: 60 Description: An introduction to machine learning methods (e.g. Artificial Neural Networks, XGBoost) and applications for snow and hydrology
Details
### Learning Outcomes * Understanding conditions under which using Neural Networks is appropriate * Learn how to prepare data for a NN workflow * Learn how to successfully train and apply a simple NN using PyTorch and save the weights for another application ### People Developing the Tutorial (content creation, helpers, teachers) ### Summary Description ### Dependencies (things people should know in advance of the tutorial) * understanding of Python for data cleaning (esp. Pandas, numpy) * understanding of ML workflow, concepts of training data, hyperparameter ### Technical Needs (GPUs? Large file storage? Unique libraries?)