One Algorithm to Secure Them All
AimBrain aims to secure consumer mobile devices and applications by developing algorithms that learn the user’s biometric model. The core technology is fuelled by the recent advances in deep learning. In the talk we’ll discuss what are the key parts of deep learning that enable us to build our platform. One key ingredient is the ability of deep nets to learn features and form representations while being trained under objectives that fit the task, rather than fitting the task to the algorithm. Composing appropriate objectives allows us to learn global latent factors that are well behaved locally, alleviating the database imbalances. The effectiveness of deep learning permits us to unify AimBrain’s products under a single learning paradigm which in turn allows us to focus our research and simplify production deployments.
Alesis Novik is the CTO and co-founder of AimBrain. With the vision to provide the market with a smart and transparent biometric experience, Alesis has helped to guide AimBrain to its unique position as the world's only multi-module, biometric security solution. Alesis holds degrees from Vilnius University and the University of Edinburgh. During the course of his academic career, Alesis participated in Google’s Summer of Code Program. More recently, Alesis has spent time at CERN and at Level E Capital. Additionally, Alesis has completed three years of PhD work.