Knowledge Distillation, a Methodology for Compressing Neural Networks

Author(s): Behzad Benam

Originally published on Towards AI.

Teacher-student architecture to create a smaller model for embedded systems

This member-only story is on us. Upgrade to access all of Medium.

Photo by Sharon McCutcheon on Unsplash

Large neural networks are complex and have a high potential to perform more complex tasks. But unfortunately, it is not easy to validate such large networks, and therefore there is no guarantee that we use the total capacity of the neural network. In contrast, smaller neural networks are easier to validate and deploy.

Knowledge distillation is a process of transferring knowledge from a large neural network or set of networks to a smaller one with an acceptable reduction in accuracy loss. This process helps compress… Read the full blog for free on Medium.

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI