🧑‍🎓About Dataset IP

In the current Web3 space, transactions are primarily focused on on-chain data, such as trading ERC20, ERC721, and ERC1155 assets. However, real-world assets go beyond just on-chain assets. Many online or offline assets that are tokenized have yet to be effectively showcased and traded. For instance, training AI models requires large datasets of text, images, videos, audio, and 3D data, which are too large to be fully stored on-chain. Decentralized storage still heavily relies on off-chain servers.

Traditional Web2 data trading depends heavily on trust between parties. Data buyers often need to verify samples, but it’s difficult to ensure that the provided samples aren’t cherry-picked by the seller. Even if the samples pass verification, it remains challenging to validate the entirety of the data. This trust gap makes it hard for data buyers and sellers to engage in reliable transactions.

Moreover, it’s often difficult for data buyers to confirm the full ownership of datasets. Sellers may download data from other platforms or resell copyrighted material without proper authorization. This leads to situations where the original owners of the data are unaware of its misuse and don’t receive any proceeds from the sale.

DeData transforms AI-generated data and application-created intellectual property into tokenized IP assets. Anyone can create their own datasets or AI application sets, and others can pay to use these IP assets. The IP owners will receive royalties for their work, and data buyers can pay for proper copyright and licensing. The use cases for dataset IP are as follows:

Our Solution

DeData offers a decentralized data marketplace, DataBazaar, which provides a trustless platform for data transactions through smart contracts. The platform enables secure, verifiable transactions between data owners and buyers. Random sampling of the data ensures that non-sampled data remains untampered and that the sampling process is fair. This guarantees both data security and the credibility of the sampling process.

When creating datasets on DeData, providers must set copyright fees, register IP rights, and stake assets as collateral for these rights. DeData also offers a dispute resolution mechanism. Anyone can raise a dispute against a dataset’s IP. If disputes accumulate, the dataset will be flagged as questionable or disputed, and trading will be halted. If the investigation reveals malpractice, the dataset provider’s staked assets will be forfeited.