FLARE Webinar 2: Federated Learning Applications: Healthcare and Finance

This is Part 2 of the FLARE quarterly webinar, focused on healthcare and financial applications
Agenda:
9:30 am PDT to 9:35 am PDT people join online.
9:35 am PDT to 11:00 am PDT Webinar and Q&A
11:00 am PDT closing

Join by:
https://teams.microsoft.com/l/meetup-join/19%3ameeting_YWVmOGY4YjgtM2Y5OS00ZTg2LWI2MjUtNDg5ODZiZmEzOWJj%40thread.v2/0?context=%7b%22Tid%22%3a%2243083d15-7273-40c1-b7db-39efd9ccc17a%22%2c%22Oid%22%3a%223f4ac28c-b68d-44d0-90db-dac09e640b21%22%7d

Microsoft Teams
Join the meeting now
Meeting ID: 215 975 200 476 1
Passcode: H9cB3vK2

Federated Learning Applications: Healthcare and Finance

Federated Learning (FL) enables decentralized AI model development while preserving privacy. NVIDIA FLARE offers tools for collaborative model training across institutions without sharing raw data. Two of the application fields are healthcare and finance. In healthcare, it enhances diagnostic accuracy by training models on sensitive patient data and enables more effective drug discovery by fine-tune foundation models with data from diverse sources. In finance, it aids in fraud detection and risk assessment through secure model training across organizations. NVIDIA FLARE supports large-scale projects, secure and efficient data exchange, model optimization, and performance monitoring for real world federated studies.
This Webinar will cover how NVIDIA FLARE is applied in these two field, with details in 1) AI models, including basic machine learning models like SVM, tree-based models like XGBoost, and deep learning models, 2) data security concerns under different collaboration modes and assumptions, 3) solutions by NVIDIA FLARE functionalities to the concerns. In particular, we will discuss healthcare use cases using both CNN models for processing medical image analysis tasks, and language models for drug discovery; and for financial applications, we will first visit how to adapt commonly-used traditional machine learning models to a federated paradigm, and then go into details of the use of XGBoost model with advanced security features. We showcase the potential to address critical challenges in distributed AI development under data sharing constraints.