AIS RESEARCH

Advancing Machine Learning

Challenge

Research scientists and engineers at AIS developed an entirely new framework for biometric verification that out-performs modern state-of-the-art methods.

The group was inspired by Deep Vectors (D-Vectors), an approach used for speaker verification. AIS scientists generalized this approach to perform verification using any biometric data source. The result was a framework that performs with higher accuracies and requires less data.

The group’s paper, “Beyond Speech: Generalizing D-Vectors for Biometric Verification” was one of only 1,147 papers selected from more than 7,700 submissions for the 33rd Annual AAAI Conference.

Key Insights:

This framework requires significantly less data than traditional approaches and performs with higher accuracies enabling practical applications
Multimodal: Generalizable framework that can extract discriminative patterns from multiple sensors and datatypes. Ideal for continuous authentication applications
Robust and practical keystroke verification: Authenticate users with just a tweet

Read the Full Document

Ready to Get Started?

Reach out to talk to one of our experts and learn more about our research initiatives.