Skip to content
Mike Pollard edited this page Nov 9, 2021 · 192 revisions
Clone this wiki locally

logo-name

ABOUT US

We created Private Identity to solve cryptography’s “holy grail,” fully homomorphic encryption. Our small team of ML engineers solved FHE in 2018 and was granted cryptography patents for FHE in 2019, 2020 and 2021. FHE conceals the input data, output data and the occurrence of the search itself.

Private Identity’s FHE algorithm is a 1-way cryptographic hash that irreversibly encrypts biometric data while enabling encrypted match and encrypted search operations on the encrypted dataset. Operating in this encrypted space fully mitigates the regulatory and legal risk of biometric data. Indeed, the GDPR, CCPA, BIPA and HIPAA privacy laws specifically state that they do not apply to anonymized data. Building on this, the new IEEE 2410-2021 Standard for Biometric Privacy provides that conforming FHE systems are, “exempt from GDPR, CCPA, BIPA and HIPAA privacy law obligations.”

In addition to privacy, FHE provides extreme accuracy, speed, security and efficiency benefits. We look forward to meeting you and supporting your cybersecurity mission and goals.

Technical Overview

Private ID® is a fast, accurate, private, efficient and scalable 1:1 and 1:n face, voice and fingerprint biometric security solution. The recognition algorithm maintains full privacy by operating entirely in the encrypted space, returns 1:n identity in polynomial time, provides 99.71% or greater accuracy across each racial subgroup (Black, Hispanic, Asian, Indian, White, Other), and easily deploys to any browser, mobile or embedded device, platform, cloud, cloud function or non-networked server using W3C WebAssembly (Wasm) or c++ sharable objects (.so). Private Identity’s high-throughput recognition algorithm consists of an ensemble of helper networks (DNNs) that validate each biometric prior to prediction or enrollment. This allows for attended or unattended operation and protects the integrity of the identity system. Each of these helper networks is an ML inference of a pre-trained DNN that locates facial landmarks, detects eyes open/closed, glasses on/off, mask on/off, image or video presentation (spoof) attack, and then crops, aligns and augments the valid biometric. FHE transformation is performed by an ML inference of a pre-trained convolutional neural network (CNN).

LEADERSHIP

Mike Pollard, CEO
Mike is an entrepreneur experienced in high-growth technology ventures. Prior to leading Private Identity, Mike served as VP and GM at Thomson Reuters, Executive VP and co-Founder of Discovery Logic, CEO of thinkXML and CEO of Science Management Corp. Mike has authored papers and patents in biometrics, AI/ML, big data and cyber security.

Scott Streit, CTO
Scott is a data scientist and inventor focused on cryptography, biometrics, AI/ML and cyber security. In addition to leading technology at Private Identity, Scott also serves as Chair of Biometric Security for IEEE (since 2013) and chair of the IEEE P2410 Standard for Biometric Privacy Working Group. Prior to co-founding Private Identity, Scott served as CTO of Hoyos Labs (now Veridium). Prior to Hoyos, he supported US national security agencies for 26 years, the last four of which he served in the role of chief scientist. Mr. Streit has authored patents and papers in machine learning, biometrics and authentication.