
Understanding Hyperparameters in Machine Learning
A stir-fry analogy
Organisation: TAFE NSW - Institute of Applied Technology Digital
Target Audience: Learners of Machine Learning Course
Platforms: Public Speeches
Ever wondered how large language models like chat GPT are built? The magic behind these lie in machine learning models, and one of the key elements in the process of building these models are hyper parameters. We were engaged to create a fun and engaging animation that explains hyper parameters and their crucial role in building machine learning models. We were given a script that drew the similarities between using hyper parameters in developing a machine learning model and the key variables of wok temperature, stirring speed and ingredient measurements when cooking a stir-fry.
This video was designed to sit within a larger online course about hyper parameters. As hyper parameters can be quite a complex topic we wanted to make sure the message was simple and clear. We used simple iconography and graphics to represent the concepts and we introduced a talking robot to guide you through some of the key concepts.
We wanted to give the audience something fun bright and futuristic to enjoy. So we decided to give the animation a retro futuristic vibe. A kitchen in a futuristic building with a floating 10 burner stove kitchen and a retro futuristic wooden spoon, for those who want the convenience of 21st century automation but with the feeling of a traditional 20th century cooking implement. We hope you enjoy the final flavour as much as we did making it.