[Edge & TinyML Track]: Model Compression Techniques for Edge AI

Model Compression is paramount in the world of Edge AI, as it is key to enhancing the performance and efficiency of AI models on edge devices. This talk will highlight the key drivers behind the increasing necessity for model compression, the essential evaluation metrics, and a range of vital techniques involved in the compression process. We will then delve into notable innovations in the field, illustrated by a case study on the Jabra Panacast20, to demonstrate the real-world applications and benefits of these techniques. The session will wrap up with a summary and a Q&A segment, equipping attendees with the knowledge and tools needed to optimize AI models for edge deployment.