2021 - 2022 · Protopia AI
Privacy-Preserving Deep Learning
Protopia generates curated noise around the source signal, safeguarding data from unauthorized inference.
Illustration - our technology on a mask-detection task over baby faces: Protopia blog.
Demo video - our technology on a face-detection task: Protopia blog.
Related: NetApp - Responsible AI and ethics by design.
Work at Protopia AI (2021.01 - 2022.08 · NeurIPS-D ‘21)
- Development of Protopia AI’s noise-training technology for inference privacy.
- Development of an internal framework and system for noise-training privacy.
- Research on face recognition with privacy preservation via noise training.
- Development of a scalable training and workflow-management system for noise training.