Research & Development

Uses the power of natural selection to generate transparent, high-performance algorithms based on modular computer vision functions

• Few training images required
• Transparent, interpretable, and verifiable by humans
• Light, fast, and sustainable

Evolutionary approaches are harnessed to streamline and refine deep neural networks — yielding lighter, faster models with powerful performance

• Speed and adaptability of evolutionary computing
• Power and performance of deep learning

Generates powerful, high-performance solutions using state-of-the-art deep neural networks

• Matches performance of human experts
• Robust and well-validated models
• Requires large training datasets
• Substantial computational requirements