Welcome to the 1st FARADISA Workshop!
It is a true pleasure to welcome you to the 1st FARADISA Workshop, held at Universitas Yarsi & Online (Hybrid). The rapid integration of Artificial Intelligence (AI) into healthcare spanning from genomic sequencing and digital pathology to clinical imaging is transforming the national landscape. However, building reliable AI models requires massive amounts of diverse clinical data, which creates a critical bottleneck: strict patient privacy regulations and institutional data sovereignty concerns make it nearly impossible, and highly risky, for hospitals to share or move sensitive patient records into centralized data lakes.
To overcome this, the FARADISA (Federated Artificial Intelligence Research Assembly for Data Interoperability and Secure Analytics) consortium has been initiated. This workshop introduces Federated Learning (FL)—a decentralized machine learning approach that allows AI models to learn from hospital data without ever requiring the raw, sensitive data to leave the hospital’s local servers. Our theme is “Federated Learning in Healthcare: From Data Privacy to Joint Publication”.
By bringing together leading universities, national referral hospitals, and pioneering industry partners who share a common vision for data privacy and interoperability, this workshop aims to break down data silos and accelerate secure AI translation into real-world clinical workflows.
Academic Initiators
- Universitas Yarsi (Indonesia)
- Kansai University (Japan)
- Kasetsart University (Thailand)
Day 1: Health Data Fundamentals & Intro to AI/Machine Learning
August 10, 2026
Lead Facilitators: Chandra Prasetyo Utomo & Muhamad Fathurahman, Universitas Yarsi
Focus: Aligning baseline knowledge. Students will learn about the complexity of hospital data, while the Hospital IT Team will learn the basics of AI/ML modeling for healthcare.
08.30 – 10.00 | Session 1 (Theory)
Healthcare Data Ecosystem & Ethics
- Characteristics of Clinical Data (EMR, HL7 FHIR, Imaging) and Genomic Data.
- Indonesia’s Personal Data Protection (PDP) Law and the challenges of cross-institutional Data Sharing.
10.15 – 12.00 | Session 2 (Theory)
Introduction to Machine Learning in Healthcare
- What is Supervised, Unsupervised, and Deep Learning?
- Why do traditional (centralized) AI models fail in hospital settings? (Privacy constraints & Siloed Data).
13.00 – 15.00 | Session 3 (Hands-on Lab 1)
Python Crash Course & Data Preprocessing
- Basic medical data cleaning (handling missing values, image resizing, data standardization).
15.15 – 16.30 | Session 4 (Hands-on Lab 2)
Building a Baseline Model (Centralized)
- Practical session: Building a basic AI/ML model (e.g., tumor classification from tabular/image data).
- Recording this model’s accuracy as the baseline for comparison.
Day 2: The Core of Federated Learning (The “How-To”)
August 11, 2026
Lead Facilitators: Kohei Ichikawa & Kundjanasith Thonglek
Focus: Introducing Distributed Computing paradigms and shifting from “sharing raw data” to “sharing model weights.”
08.30 – 10.00 | Session 5 (Theory)
Federated Learning & Distributed Computing Architecture
- The concept of Server (Aggregator) vs. Client (Hospital) in an HPC environment.
- The Federated Averaging (FedAvg) algorithm explained in simple terms.
10.15 – 12.00 | Session 6 (Hands-on Lab 3)
Setting Up a Federated Learning Simulation
- Introduction to an FL Framework.
- Running an FL simulation locally within a single machine.
13.00 – 15.00 | Session 7 (Hands-on Lab 4)
Multi-Client Federated Learning (Roleplay)
- The most crucial and interactive session!
- Participants are assigned roles simulating real institutions (e.g., Server: BGSi/Industry; Client 1: RSKD; Client 2: RSCM).
- They will train the model separately on their respective laptops but communicate to update the global model on the Server.
15.15 – 16.30 | Session 8 (Theory & Discussion)
Non-IID Data Challenges in the Real World
- What happens if data distribution significantly differs across centers? (Introduction to Data Heterogeneity).
Day 3: Advanced Capabilities, Genomics, and Joint Publication
August 12, 2026
Joint Facilitation: International Experts & Yarsi
Focus: Applying FL to real-world cases (Genomics/Imaging) and formulating collaborative research for international journals.
08.30 – 10.00 | Session 9 (Theory)
Federated Learning for Genomic and Medical Imaging Data
- Implementing FL on large-scale data (e.g., whole slide images for pathology, pediatric clinical images, or genomic variants).
- Cross-institutional network infrastructure and latency issues.
10.15 – 12.00 | Session 10 (Execution Workshop)
Mini Capstone Project
- Collaboration in mixed groups (Lecturers + Students + Hospital IT + Industry).
- Task: Modify the Day 2 lab code to achieve a target accuracy using clinical/genomic dummy datasets.
13.00 – 14.30 | Session 11 (Strategic Material)
From Practical Labs to Q1/Q2 Joint Publications
- Deconstructing the structure of FL papers in top-tier journals.
- Establishing future authorship roles based on institutional strengths.
14.45 – 16.00 | Session 12 (Closing & Agreement)
Launching FARADISA Working Groups
- Pitching joint research ideas from each group based on existing/new projects.
- Agreeing on a post-workshop research sprint schedule.
