Prospects for Combining Distributed Ledger Technology (Blockchain Technology) and the Process of Federal Machine Training of Artificial Intelligence (AI) in the Form of a Unified Information-Protected Structure
The paper addresses the challenge of ensuring the integrity of AI training data by integrating distributed ledger systems with blockchain technology, leveraging cryptographic methods and federal AI training techniques. Ensuring the confidentiality of data and its protection from falsification, modif...
Сохранить в:
| Главные авторы: | , |
|---|---|
| 格式: | Статья |
| 語言: | English |
| 出版: |
Springer Nature
2025
|
| 主題: | |
| 在線閱讀: | https://dspace.ncfu.ru/handle/123456789/30673 |
| 標簽: |
添加標簽
沒有標簽, 成為第一個標記此記錄!
|
| 總結: | The paper addresses the challenge of ensuring the integrity of AI training data by integrating distributed ledger systems with blockchain technology, leveraging cryptographic methods and federal AI training techniques. Ensuring the confidentiality of data and its protection from falsification, modification, and destruction is currently an important task in producing AI systems. Thus, the solution to the above-listed tasks in terms of ensuring information security of data in all their variety is rather relevant. The paper identifies five key “ifs” that define the properties a distributed ledger system using blockchain technology must possess. From this, five imperative principles, with unique characteristics not found in existing systems, have been developed for designing such systems. The inclusion of timestamps in the protocols of a notary cryptographer, in the form of an additional hash linked to the main hash that connects blocks within a peer-to-peer network, is considered valuable for transparent control of document creation time by all participants in the blockchain network. The use of timestamps alongside traditional cryptographic authentication and identification technologies, which ensure subscriber trust, is shown to be beneficial. This method enables all network participants to verify the “freshness” of the information provided by other subscribers. |
|---|