What is Raven Protocol?

Source: Deep Learning on Medium

Raven Protocol in a short span has become a very well known name in the crypto space for its innovative use of blockchain technology for AI and also for conducting the first ever Initial Decentralised offering. The project gained a lot of traction and community support in the last few months and has successfully raised $6.1Million USD from multiple investors. The project has also successfully raised $500,000 USD through the IDO, which lasted a few seconds.

So, what is Raven Protocol?

Well, as stated on their website, “Raven Protocol is a decentralized and distributed deep-learning training protocol which provides cost-efficient and faster training of deep neural networks by utilising the compute resources in the network”.

What does that mean for a normal individual?

In layman’s terms Raven Protocol is building a decentralised protocol which would be using the computation power from systems to power the training of deep neural networks. It can also be called as a super-computer, which is powered by multiple systems(Contributors) on the network.

There might be few questions in your mind as, what is deep neural network? What is its use? How does that work?


Deep neural network is a part of Machine learning. Machine learning uses a set of patterns or inference to complete tasks. Which would mean that you do not have to give any instructions to the system. The system follows certain algorithms and statistical models to perform the tasks.

Deep learning is a subset of Machine learning, which is capable of learning unsupervised from data that is unstructured or un-labeled. Deep learning can be used in multiple ways such as translations, language recognition, object classification and many more. Deep learning helps computers to recognise things by themselves without need of human instructions or interactions.

One such example would be automated video suggestions on streaming websites based on the multiple parameters such as categories, time, language etc.. The system could also provide videos depending on the mood of the person or just simply state sentiment on how the day for the person has been depending on his past history.

The current servers and systems take lot of time and resources to complete neural trainings which could span days depending on the server’s capabilities at higher costs.

Raven protocol is building a distributed and decentralised network where anyone can join the network and contribute. The network would be capable of processing data which could take days on traditional servers to hours or minutes. Any one who joins the network/participates by providing computational power would be rewarded with the Raven tokens for their contribution. The contributor can allocate free computational power to the network and let the network do the magic and reward them with the Raven tokens.