。 A medical team said that the blockchain is useful for cancer researchers to share information for their AI systems without the help of the central coordinator. The team specifically mentioned the use of smart contracts on Ethereum for this purpose; Ethereum's smart contract enables three different research teams to update their AI models simultaneously without going through a central authority. The artificial intelligence model itself is used to predict the emergence of cancer cells in vivo. The following is a detailed content sharing.

A medical team proposed in a paper this year that blockchain is useful for cancer researchers. A research paper entitled Swarm learning for decentralized artificial intelligence in cancer pathology, published in Nature Medicine in April, was written by 27 different authors. In a footnote of the paper, it was pointed out that the team began to use the Ethereum network for cancer experiments.

According to this paper, artificial intelligence (AI) can help predict the occurrence of cancer cells in patients by extracting information about the shape and size of cells invisible to the human eye. However, from the perspective of data collection, the large data sets necessary to run such AI systems face "practical, ethical and legal obstacles", especially if data are shared between countries.

One way to solve this problem is to use joint learning (FL), which does not require researchers to share their data, but only the AI model weights (or parameters) they trained locally. The problem is that this system relies on a centralized coordinator, who essentially combines all the model weights together, and then fully controls the research project and its commercial development.

On the contrary, the team pointed out that the use of group learning (SL) is increasing, which is a system using blockchain technology to avoid transferring power to a centralized entity. In other words, SL enables teams to share their AI model weights while keeping all contributors at the same level, which makes it easier for more people to cooperate. In turn, it provides more data for AI models, making them more powerful.

In particular, the research team pointed out that it used smart contracts on Ethereum so that three independent computers could synchronize their AI model weights at a specified time. In fact, all three partners have updated the AI model at the same time, without the help of the coordinator, who will manually merge the model parameters. "In this setting," the paper said, "the blockchain maintains the global state information of the model."

The research paper finds that the AI system born from this setting is superior to the AI model trained locally, and its performance is comparable to that of other models trained with consolidated data sets (and the data efficiency of this technology is higher). As Ari GoldNFT, a medical professional, explained when pointing out this article on Twitter, "A hospital in New York can communicate with a hospital in Los Angeles through a node."

epilogue

This is an important message for general encryption technologies, especially smart contract platforms. So far, blockchain has proved to be very useful in the financial field, but critics and enthusiasts have condemned the lack of adoption of this technology in other fields. Vitalik, the creator of Ethereum Butlin said in August that cryptocurrencies need to be "turned into something useful" in the next decade. It is difficult to find more valuable use cases than in the medical field.