- A new, IOTA-based medical diagnostic AI reaches 96% accuracy without data exchange.
- Data protection is achieved by secure digital identities with integrated governance.
A new research work that makes waves in the academic world shows that the IOTA infrastructure is driving a data protection-oriented revolution of artificial intelligence in healthcare.
The System Works by being able to train AI models together without having to transmit sensitive data directly with each other. Although the paper was written anonymously, the work is largely considered the product of the productive DAC2025 author Arpita Sarker.
#IOTA is at the core of one of the most promising AI papers of the year — a groundbreaking system for training artificial intelligence models collaboratively, privately, and without intermediaries.
The paper was published anonymously, but all signs point to Arpita Sarker…
– Salima (@alamasburg) July 20, 2025
The system includes local training in every organization, After which AI upgrades About Ipfs and Tangle from Iota be provided. Thanks to the decentralized Frame the participants can Provision Vote over each upgrade. Each process is recorded unchangeable, which means that the data integrity is preserved and fees and central intermediaries become superfluous.
The first results are amazing. A with the system trained AI model achieved an accuracy of 96% in the Detection of Diseases from clinical texts. Another classifier arranged The reports automatically After the official Diagnostic brains a.
A third system was on it trainedon the basis of data generated by voice models such as Llama and biogpt, to answer questions about healthcare, which in turn shows the effectiveness of this decentralized learning paradigm.
IOTA and that Decentralized identity management
There is one Further work that IOTA’s potential shows in the administration of decentralized identifiers (DIDS). While the vast majority of blockchain networks still In the experimental phase of researching the field, Iota has real resources for the joint management of identities through accounts with several signatures and weighted votes.
– Salima (@alamasburg) July 20, 2025
Technologies enable it Groups like companies or Daos , Digital identities without central authoritiescooperative to create and manage. With built -in functions that support group management, can The parties Common identity documents secure And in real timechange.
This concept was in dem Research paper from the researchers of the Technical University of Berlin Carlo Segat, Sandro Garzon Rodrigueand Axel Kupperfurther developed. It suggests the integration of governance rules directly into DDOS (Decentralized Document Objects) in order to enable programmable decision-making.
In addition, improvements such as verifiable authorization evidence, token authorization and a greater Flexibility in the vote between different ledger systemsproposed .
Research signals great potential for future digital infrastructure
Both studies emphasize The growing importance of IOTA in decentralized secure systems. From support in the development of AI models in extremely sensitive sectors such as medicine to the redefinition of the way, such as identities In the online world manageIOTAS proves their potential free of charge, latency infrastructure.
No Comments