Blockchain Ireland Tech Working Group - Tech Talks - ACR Platform
Title: Blockchain Ireland Technical Working Group: ACR Platform Deep Dive
Date: October 7th
Location: Online
About the Event:
As co-chair of the Blockchain Ireland Technical Working Group, I am excited to invite you to our upcoming online session exploring the intersection of distributed ledger technology and healthcare AI.
We will be taking an in-depth look at the ACR Platform (AI Cancer Research Platform), an innovative solution designed to demonstrate a practical role for DLT below the application layer.
What is the ACR Platform?
The ACR Platform is an AI-assisted clinical decision support (CDS) platform initially developed for breast-cancer decision support and built to extend to other clinical domains.
It moves away from black-box diagnoses by providing explainable decision support, allowing authorized clinicians to inspect clinical facts, reasoning paths, and governing rules.
The system combines clinician-validated semantic knowledge, deterministic reasoning, optional probabilistic assistance, and distributed AI governance.
How It Works & Core Architecture
The platform operates on a strict design principle of "Data Stays. Rules Travel," ensuring that patient data remains securely under the control of its custodian institution.
Patient records are never placed on-chain; instead, the DLT network is utilized as the trust, provenance, and consensus layer for multi-institution AI governance.
The technical stack features an OWL clinical ontology utilizing SWRL rules and Openllet, alongside a Bayesian layer for probabilistic assistance.
The architecture also incorporates federated and reinforcement learning, allowing AI agents to submit governance recommendations that authorized institutions can review.
The platform and its companion mobile applications were built using "Kraken’s Loop," a finite, evidence-driven control method for autonomous AI agent engineering that enforces strict stop gates and human release authority.
About the Speaker:
The platform will be presented by its developer, Kraken YU.
Kraken is a senior software engineer holding an MSc in Advanced Software Engineering from UCD, with a career spanning object-to-agent-oriented systems, blockchain, AI, decentralisation, and tokenisation.