Seeing Green

Two Topics: One on making some green and the other using "green" AI.

Financial Machine Learning II: Pilar Orellana, In this note, we showcase optimization techniques for an artificial neural network with respect to stock price prediction. If time permits, we'll discuss the role of the Black-Scholes model, option strikes, and second and third order Greeks in Machine Learning.

Green ML/AI: Machine learning and AI tend to consume a lot of energy during training. It is not unusual to have multiple computers running for multiple days to train even a simple model. Green Algorithm on the other hand are smaller statistical models that take only a fraction of the time to train and can similar and sometimes better results. This talk will go over one of these "green" algorithms.

Future Topic ideas:

• MLFlow: Brief overview what it is and why you should switch.
• Multi Arm Bandit: Why and how you should be using this to design your experiments.
• Data: Your algorithm is only as good as your data. DVA promises to help you keep track of any changes you made and when to your data. Does it live up to the hype? Is it really git for your data?
• Training Techniques: Have you start using the classroom technique. What training techniques do you utilize.