GalsenAI invite you to a very special event: a workshop about Image classification with Transfer Learning using Tensorflow and Deep Learning applied to audio.
One of the most powerful ideas in Deep Learning is that we can take the knowledge that a neural network has learned from one task and apply that knowledge to another task. This is called Transfer Learning. Transfer learning is adapting a model trained for one purpose to be used for another. Transfer learning works surprisingly well for many problems, thanks to the features learned by deep neural networks. In the first place of this workshop, Thierno Ibrahima Diop, lead data scientist at baamtu, will cover the basic methodology of transfer learning and showcase the results in the context of image classification.
While working with images is quite common in Deep Learning, dealing with audio raises a whole new set of challenges and techniques that are needed to be considered. So in the second place, Sam Witteveen, a Google Developers Expert in Machine Learning and Co-founder & CEO Red Dragon AI, will look at the basics of working with audio and building Deep Learning models to do a variety of tasks such as audio classification, tagging, and speech recognition.
We will be building an audio classification model from scratch and along the way look at the common pipelines and libraries that are used as well as the types of Deep Learning architectures which get the best results for this kind of task.
After going through some basic examples in code, Sam will then look at how models such as speech to text models work and how they are combined elements from NLP to make a production-grade system in products such as Google Home and Alexa.
Don’t miss this event !