The implementation of the NeuroSeed platform will lead to the rapid development of the entire set of machine learning technologies. It will help to reduce the costs of development, mass deployment and significant increase in the efficiency of derivative systems. The concept of the rationality of using several machine learning models merged and further pre-trained is proposed and approved. A common limiting factor in the development and implementation of such systems was the lack of reliable technologies that could provide decentralized digital reliability for the machine learning models and data sources. This technology has become a blockchain.
Use of Blockchain on the platform
Blockchain is used for storing the hashes of Big Data and Machine Learning models together with their authorship metadata. The metadata is used by search algorithms when Customer/Validator/Distributor request a dataset or ML model. Hash is also used to identify and extract the target item from the distributed data storage. The unrivaled advantages of the platform are the transparency and security of payments, datasets and ML models that are provided by cryptographic information security technologies.
Motivation and validation
To improve the quality, all data and decisions are validated by external experts. Companies and specialists with domain expertise can become platform validators. There is a special motivation system for validators. The reputation and payments for their verification job will grow within the system accordingly to the quality of their work. Incorrect datasets and ML models validation affects the internal and external reputation of the validator.
