Raven Protocol is a decentralized distributed deep-learning training protocol that will aim to provide cost-efficient and faster training of deep neural networks.
Raven Protocol’s specific use case is to perform AI training where speed is the key. We’re taking a 1M image dataset that takes 2-3 weeks to train on AWS down to 2-3 hours on Raven. AI companies will be able to train models better and faster.
That’s just the beginning. The vision of Raven Protocol is to pull the power of AI/ML out of the hands of mega-corporations like Amazon/Facebook/Google and put it into the hands of the community. Only then can we unlock value and create things we never would have imagined with the power of AI. Large companies have an unfair advantage in terms of AI/ML resources and have no reason to share it with the rest of the world.
AI and Machine Learning needs to be decentralized. Blockchain is going to transform the entire AI industry. It enables three key ingredients that never have been possible before:
- Blockchain enables the sharing of training data.
- Blockchain enables the sharing of research algorithms.
- Blockchain enables the sharing of compute resources.
At the core of this is Raven Protocol which focuses on compute resources to speed up AI training. That foundational layer will allow you to throw massive datasets and any machine learning algorithms you want to train with. Raven will give you the power to train AI models better and faster.