FHE Fully Homomorphic Encryption, although it is difficult to mature and land for a short time and be applied, it is an extension and complement to ZKP technology.

Author: Haotian

Vitalik's article on Fully Homomorphic Encryption (FHE) has once again sparked everyone's exploration and imagination of new encryption technologies. In my opinion, FHE Fully Homomorphic Encryption indeed expands the imagination space more than ZKP technology, and can help more applications of AI+Crypto to land. How should we understand this?

1) Definition: FHE Fully Homomorphic Encryption can perform operations on encrypted data of specific forms without worrying about exposing data and privacy. In contrast, ZKP can only solve the problem of consistent transmission of data in encrypted state. The receiving party can only verify that the submitted data from the submitting party is authentic, which is a point-to-point encrypted transmission scheme. Fully homomorphic encryption does not limit the range of entities performing operations, and can therefore be considered as a many-to-many encryption operation scheme.

2) How it works: Traditional computer operations are performed on plaintext data. If it is encrypted data, it needs to be decrypted before calculation, which inevitably exposes private data. Homomorphic encryption constructs a special encryption scheme that can perform "homomorphic" transformations on ciphertext, so that the results of the operation are the same as performing the operation on plaintext. In a homomorphic encryption system, plaintext addition is equivalent to ciphertext multiplication (a rule), so if you want to perform addition on plaintext data, you only need to perform multiplication calculation on ciphertext (equivalence).

In short, homomorphic encryption uses special homomorphic transformations to allow data to be operated on in ciphertext state and obtain the same results as plaintext, as long as the homomorphic correspondence property of the operation rules is ensured.

3) Application scenarios: In the traditional Internet field, FHE Fully Homomorphic Encryption can be widely applied in cloud storage, biometrics, healthcare, finance, advertising delivery, genetic sequencing, and many other fields. Taking biometrics as an example, personal biometric data such as fingerprints, irises, and facial features are sensitive data. Using FHE technology, these data can be compared and verified in ciphertext state on the server. Similarly, in the healthcare field, years of fragmented data can be broken using FHE, allowing different healthcare structures to conduct joint analysis and modeling without sharing original data.

In the Crypto field, the application space of FHE can also involve games, DAO voting governance, MEV protection, privacy transactions, regulatory compliance, and many other scenarios that require strong privacy. Taking the game scenario as an example: the platform can drive the game to be more fair without peeking at the cards in the players' hands.

Similarly, in the case of DAO voting, whales participating in voting governance can generate voting results through calculations without exposing addresses and voting quantities. In addition, users can submit encrypted transactions to the Mempool to avoid exposing private messages such as target addresses and transfer amounts. For example, in a regulatory scenario, the government can monitor fund pools and separate assets from blacklisted addresses without viewing the privacy data of legitimate transactions.

4) Shortcomings: It is worth noting that the computing environment for performing regular operations on plaintext in a computer is often quite complex. In addition to addition, subtraction, multiplication, and division, there are also conditional loops, logical gate judgments, and so on. However, semi-homomorphic encryption and fully homomorphic encryption currently can only be quickly expanded on addition and multiplication, and more complex operations require combination and superposition, which correspondingly increases the computational power requirements.

Therefore, theoretically, fully homomorphic encryption can support arbitrary calculations, but due to performance bottlenecks and algorithm characteristics, the types and complexity of homomorphic calculations that can be efficiently executed at present are very limited. Complex operations will require significant computational power consumption. Therefore, the process of landing fully homomorphic encryption technology is actually a development process of algorithm optimization and control of computational power costs, especially focusing on performance after hardware acceleration and computational power enhancement.

Above

In my opinion, although it is difficult for Fully Homomorphic Encryption to mature and land for a short time and be applied, as an extension and complement to ZKP technology, it can provide great value for AI large model privacy computation, AI data joint modeling, AI collaborative training, Crypto privacy-compliant transactions, and extension of Crypto scenarios.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。