For Cryin' Out Loud! How Neural Networks Made Me a Better Parent

3:15pm - 3:40pm on Friday, October 4 in PennTop South

Andrew Hao


Parenthood is hard! In the months after my son arrived, I found myself struggling through depression, especially so during long episodes of infant crying. So I coped by beginning a journey to build an AI baby monitor, training a TensorFlow model to recognize and quantify my little one’s cries. But this process was fraught with confusion - I had never worked with these models before. How do they work? And more importantly, how will knowing all this help me survive the travails of early parenthood?

Join me on a journey through the inner workings of a TensorFlow audio recognition model. Together, we’ll see how it works by walking through a CNN described from a Google research paper. We’ll explore core TensorFlow model concepts of weights, biases, and convolutions and learn how they work in the process. Even if you don’t have any experience in machine learning, you’ll walk away with key understanding of how ML models are structured so you can apply them to your next project!

Want to edit this page?