Paper Reading: Intro to Continual Deep Learning

Deep Learning models are increasingly compute and data intensive. So how can an intelligent system continually update itself in an efficient way? Unlike traditional approaches where deep learning models are periodically re-trained on the expanded dataset, continual learning enables models to iteratively adapt as new data is collected - without re-training the model from scratch!

Join this interactive session on continual learning 101 to uncover:

• How continual learning can be useful for tackling high computational costs and data-storage constraints.
• What is catastrophic forgetting, and why models struggle with stability-plasticity trade-off.
• State-of-the-art approaches to Continual Learning.
• Which metrics are used to measure continual learning performance.

This event is perfect for software engineers as well as deep learning enthusiasts and practitioners looking to continually learn more about AI ;)

Your host:

Hemang Chawla is an applied scientist in computer vision focusing on anti-counterfeiting at Scantrust. He has previous worked in the domains of robotics and mapping for ADAS, and has published several papers at top conferences such as ICRA, IROS, and WACV. Checkout his latest paper on Continual Learning.

This event is free for all. For premium access to Taro, you may use this referral link for a 20% discount.