Another AI+Crypto project with the support of three major VCs, Coinbase, Paradigm, and Polychain, was born - Vana. Simply put, Vana aims to tokenize user private data, build a distribution and incentive network for user data ownership and control rights and future economic benefits, and solve the problem of insufficient data in the training process of large AI models.
The development of AI mainly revolves around three major areas: "computing power", "algorithm" and "data". The computing power directions include io and Aethir, and the algorithm directions include Bittensor and SaharaAI. Vana is focusing on the "data" direction, which is regarded as the fuel of AI.
With the support of a huge "data source", AI can perform multimodal learning, continuous learning, self-supervised learning, etc. to enhance the application scenarios and usage scope of large AI models.
At present, AI large-scale model training faces many challenges, such as privacy and uneven data sources. There is a serious oversupply of general text-based network data, while high-quality data in specific fields (medical, legal), and real-time updated data sources (news, technology) are in serious shortage. How can we break the data silos of traditional industries, reduce data labeling costs, and effectively solve complex issues such as privacy?
After a preliminary look at Vana's technical documentation, it attempts to build:
1) Data Liquidity layer, which allows you to use data on the Vana network just like flexibly using tokens in the DeFi system;
2) Data Portability layer, which is equivalent to building an ecosystem where data providers, developers, and platforms can collaborate to promote the orderly flow of data, allowing developers to directly use data through tool interfaces, and also recording and incentivizing high-quality data contributions;
3) The data "neural network system" (Connectome) builds a distributed ledger that can record real-time data transactions in the ecosystem, as well as a set of POS consensus mechanisms to ensure that the DLP liquidity layer can operate normally and is compatible with the external EVM environment. This is the core of the Vana mainnet's ability to effectively solve AI data problems and is the key infrastructure for converting "data" into measurable value and traceable liquidity.
