8. User types

In this section we outline the different user types that NetMind power aims to serve. Given the system is based on a tripartite design, we will cover the demand side (users of the platform), the supply side (those who provide computing power to the platform) and the blockchain side (contributing to the running of the blockchain).

8.1 Demand Side

1. Machine learning practitioners

Practitioners and experts in machine learning and deep learning worldwide will be able to use our platform for training and fine-tuning machine learning models that are either proprietary or open source. Once they have trained their models, the user can host the model for inference on our platform, accessible via API, to embed their models into their applications, or for batch inference. They will also be able to make their models available for inference through the platform to the community and be compensated for doing so, without having to share or publish their models' weights.

2. Researchers

Likewise, AI researchers from universities, government research institutions, private companies and research labs alike will be able to use the training platform to train and fine-tune their machine learning models in a cost effective way to advance their research efforts. Like machine learning practitioners, researchers can also choose to make their models available for inference to third parties for a fee.

3. Developers

Software developers that aren't necessarily experts in Machine Learning and that are seeking to embed AI and machine learning capabilities into their applications will be able to use our Inference platform to do so in a cost effective and scalable way via API. They will be able to host their own proprietary models on our platform, choose from a selection of open source off-the-shelf models, or pay to use models trained and published by other users of the platform (in which case they will have to pay for both the GPU resources and the fee set by the model's owner). Additionally, the training platform is designed for these users, in that it allows for a no code way to fine tune existing open source models - the user just provides the dataset to fine tune the model on.

8.2 Supply side

1. Volunteer Computing Resource Provider

Anyone owning one or more GPUs, whether a gamer, cryptocurrency enthusiast, crypto miner for example can contribute their idle machines to the network for other users to train, fine-tune and run inference on, and be rewarded with NMT (NetMind Token). NMT is the native utility token of the NetMind Power platform and is discussed in a later section of this white paper.

8.3 Blockchain side

1. NetMind Chain Mind Node

Mind Nodes validate transactions and create blocks in the blockchain. Users can connect their machines to the network for this purpose, and are rewarded with NMT as outlined in the NetMind Chain section of this white paper.