Next meetup scheduled: Privacy-Preserving Graph ML with FHE for Collaborative Anti-Money Laundering w/ Fabrianne Effendi
on Nov 14th, 1pm CEST (Paris, FR)
🗓️ The next FHE.org meetup has been scheduled for next week, Thursday, Nov 14th at 1pm CEST (Paris, FR).
This meetup features Fabrianne Effendi, an AWS Associate Solutions Architect and recent graduate of Nanyang Technological University Singapore, presenting Privacy-Preserving Graph ML with FHE for Collaborative Anti-Money Laundering.
For more information and link to RSVP, see the event page at https://fhe.org/meetups/062.
Abstract
Fabrianne Effendi will be presenting her university final year project and research publication titled, “Privacy-Preserving Graph-Based Machine Learning with Fully Homomorphic Encryption for Collaborative Anti-Money Laundering". Co-authored with Assoc Prof Anupam Chattopadhyay, this paper has been accepted for the 14th International Conference on Security, Privacy, and Applied Cryptographic Engineering (SPACE) 2024.
With digitalization and cybercrime on the rise, anti-money laundering (AML) has become increasingly complex. Graph-based machine learning techniques have emerged as promising tools for AML detection by capturing hidden relationships within financial networks. However, data silos and strict privacy regulations hinder collaboration between financial institutions, limiting the effectiveness of AML solutions.
This presentation introduces a novel privacy-preserving approach for collaborative AML machine learning, facilitating secure data sharing across institutions and borders while preserving privacy and regulatory compliance.
Leveraging Fully Homomorphic Encryption over the Torus (TFHE), computations are directly performed on encrypted data, ensuring the confidentiality of financial data. Notably, this research explores the integration of Fully Homomorphic Encryption over the Torus (TFHE) with graph-based machine learning techniques using Zama Concrete ML. The research contributes two key privacy-preserving pipelines:
Exploration of the feasibility of developing a privacy-preserving Graph Neural Network (GNN) pipeline, optimized with techniques like quantization and pruning for FHE compatibility.
Development of a privacy-preserving graph-based XGBoost pipeline that pre-processes the data using Graph Feature Preprocessor (GFP), with experimental results demonstrating strong performance.
About the speaker
Fabrianne Effendi is an Associate Solutions Architect at Amazon Web Services. She recently graduated from Nanyang Technological University with a Double Bachelor's Degree in Computer Science and Business, specializing in Artificial Intelligence and Data Science and Analytics.
In Summer 2023, Fabrianne interned as a Software Engineer at JPMorgan Chase & Co., focusing on anti-money laundering (AML) detection in the Financial Crimes department. This experience inspired her university final year project, “Privacy-Preserving Graph-Based Machine Learning with Fully Homomorphic Encryption for Collaborative Anti-Money Laundering”, completed under the supervision and mentorship of Assoc Prof Anupam Chattopadhyay.
Never miss an update
Join the discord server to discuss FHE related topics with the community: discord.fhe.org
See you soon!
The FHE.org team