Find new｜Pastel: NFT platform with machine learning screening

"Find New" is a blockchain project observation project launched by Jinse Finance. It covers the development of projects in various fields of the industry, and the specific design includes project overview, technological progress, fundraising situation, etc., and strives to present you a collection of popular and trendy projects. .

Pastel is a decentralized cryptocurrency project for the creation, verification and exchange of provably rare digital artworks. Digital artwork refers to encrypted digital collectibles that are created in a manner similar to how traditional artists create limited-edition physical prints.

The Pastel team believes that the idea of rare digital art will change the world and revolutionize the way art is created and consumed, allowing artists and collectors to connect directly without paying huge sums to intermediaries such as Sotheby's and famous art dealers. cost.

A key feature of Pastel is that it runs on a fully native blockchain.  This ensures that Pastel is not tied to any other blockchain or network, such as is the case on Ethereum: an increase in traffic drives up transaction fees, making it economically difficult to register or purchase a piece of art.

The total market value of metaverse-related encrypted assets is US$10.4 billion: Jinse Finance reported that according to the latest data, the total market value of metaverse-related encrypted assets rebounded to US$10.4 billion, a 24-hour drop of 6.44%, and a 24-hour trading volume of US$2.13 billion. The top five NFT tokens by current market value are: Decentraland ($1.49 billion), ApeCoin ($1.24 billion), Theta Network ($1.23 billion), Axie Infinity ($1.14 billion), TheSandbox ($1.10 billion). [2022/6/22 1:24:07]

NFT platform Ucollex completes US$10 million in Series A financing, led by Animoca Brands and others: On February 24, NFT platform Ucollex completed US$10 million in Series A financing, led by Animoca Brands and MCP IPX One Fund of Japanese investment company MCP Asset Management , an undisclosed valuation. The financing will be used to scale operations, including platform and blockchain technology development, and will also help drive marketing efforts related to key intellectual property (IP) partnerships. (CoinDesk) [2022/2/24 10:12:08]

Comparison of Pastel and other platforms

Because Pastel is an independent blockchain, unlike other NFT platforms that rely on Ethereum and are hampered by network congestion, Pastel hopes to be an affordable option for any artist looking to create and sell their work. So make sure to provide rewards to node operators who are responsible for network security and become the best NFT blockchain choice for users who want to participate in the digital art ecosystem in the long term.

Bitcoin's estimated leverage ratio hits a record high: Jinse Finance reported that on January 7, according to data from the on-chain analysis resource CryptoQuant, as the price of Bitcoin continued to decline, its estimated leverage ratio had reached 0.224, a record high. Indicating that more investors are taking highly leveraged risks. [2022/1/8 8:33:31]

In addition, Pastel also has the privacy properties of ZCash and Dash and has distributed storage. Many NFT projects on Ethereum store the URI data of the token in the cloud, and in many cases it is just an AWS hosting service that needs to be paid for or "rented" for the next few years, which gives artists and collectors Home presents a serious dilemma.

To deal with the persistent storage problem, Pastel built a new distributed file storage layer. Pastel distributes "shards" across the network among masternodes, giving hundreds of network participants an incentive to keep a small portion of each artwork and "persist" or share it with the network, even in the event of a major network outage The case can eventually be rebuilt as well.

Cardano founder: End-to-end microfinance transactions will be realized in the second half of 2022: On December 26th, Cardano founder, IOHK co-founder and CEO Charles Hoskinson talked about how to End-to-end microfinance transactions on Cardano in the second half of 2022. “My goal in the second half of 2022 is to figure out how to bring all the pieces together to enable end-to-end microfinance transactions on Cardano. So a real person in Kenya or somewhere, has a blockchain identity And credit scoring, on the other side of the transaction is the stablecoin, and Cardano is the settlement rail. Peer-to-peer, person-to-person, click a button, and the loan is given to them. They pay the money back, and the money goes to the other side.”( Cryptoglobe)[2021/12/26 8:04:40]

Infringement of digital artwork is a headache in the industry, and Pastel uses deep learning to solve this problem.

Pastel uses 5 different deep learning models (VGG19, Xception, InceptionResNetV2, DenseNet 201, and InceptionV3) to create a combined image fingerprint vector for each uploaded work. Because each model is built using a different architecture, and different architectures learn different types of works in turn, by combining 5 different models with different architectures, Pastel makes technical imitation infinitely difficult .

By applying the Hoeffding algorithm and independently checking the fingerprint vectors of the created artworks by at least 3 masternodes, any new registration attempts with near-similar fingerprint images will be rejected. This protects artists and prevents malicious masternode operators from creating fakes to steal credit for others.

The details of the NSFW check are relatively straightforward, and Pastel uses a pre-trained TensorFlow model from the Yahoo! OpenNSFW project to evaluate each submitted image file. The model assigned a score between 0 (non-NSFW) and 1 (non-NSFW). If the scoring result is above a certain threshold, the master node rejects the submission as unsuitable. By keeping scores relatively low, Pastel ensures that most inappropriate content-zce is never allowed on the web, and this NSFW detection establishes a baseline confidence in the attributes (and legitimacy) of Pastel web-stored image content.

For ownership issues, Pastel has also introduced a simple legal contract that users of the Pastel Network NFT platform can sign on a special ticket (bearer copyright transfer ticket) on the Pastel Network blockchain. This ticket is only valid for Pastel NFTs for which a single, unique "copy" or digital print exists (i.e. if a user registers the NFT on Pastel and creates a limited edition of 10 digital prints), the idea is similar to a bearer bond , an unregistered debt instrument that can be redeemed like cash by the owner without showing other supporting materials.

The basic idea is that the original NFT creator agrees that if someone makes a request to prove they own the NFT, the owner of the NFT can use the Pastel_ID private key to sign the message (which only the owner of the NFT knows), then they will make the copyright official. Transfer to the person or entity making the legal claim.

However, the ownership of the copyright will be subject to the same requirements for the next owner: that is, if the other party can prove that they are the new owner of the NFT, they can request, and the copyright ownership will be transferred to someone else.

Clearly, this aspect of the system is not fully decentralized. However, the parties can enter into any contract they want, and the legal system can enforce the contract to the extent that the parties' claims are valid. In this case, the original creator of the NFT will decide in advance that they are willing to enter into this agreement with the next owner of the NFT, but only if the next owner accepts the same terms of ownership. This would create enforceable contracts, and even if lawsuits and court rulings are ultimately required, current owners of NFTs should be able to convince judges and juries to pursue their copyrights.

Tags：

NEAR