An Attention Based Approach for Automated Account Linkage in Federated Identity Management

Title:

An Attention Based Approach for Automated Account Linkage in Federated Identity Management

Abstract:

Linking digital accounts belonging to the same user has progressed from a research topic to a foundation for security, user satisfaction, and developing next-generation services. Still, few studies address account linkage in domains other than social networks. This deficiency is particularly apparent in the federated environments in academia, where network-based information and contextual data are typically unavailable.

In this paper, we propose a self-attention-based model named SmartSSO-SBERT that learns user routines and behavior during login processes to automate account linkage models for research and educational institutes. SmartSSO-SBERT generates new representations from user authentication requests in a lower-dimensional latent space where the learned structure is used to identify related accounts held by a user.

We show the trained model using a large corpus of production data consisting of more than one million samples gathered over six months from 50,000 users achieves over 98% accuracy in terms of the hit-precision.

Presenter: Shirin Dabbaghi

Location: NII 1512

Time: 2023-04-11 13:00 ~ 14:00

Other information:

Please email us to get the meeting link if you are intersted!