Valgo

Valgo is a public benefit corporation developing a risk quantification layer for physical AI, with a focus on validation, deployment, and insurance. Its work addresses the challenge of assessing and quantifying risk in autonomous and cyber-physical systems, including autonomous vehicles and robotics.

The company's research covers a range of safety validation and risk assessment methods. These include failure sampling and rare-event discovery, failure probability estimation for black-box systems, reachability analysis, runtime monitoring, and safety validation algorithms. The research also includes work on surrogate modeling for safe planning under uncertainty, probabilistic parameter estimation, and the certification of machine learning systems in safety-critical applications. Valgo has also published a tool, Human Crash Baselines, which computes human crash-rate baselines for robotaxis and autonomous trucks using specification-curve analysis.

Valgo's work is aimed at developers and researchers working on autonomous vehicles, robotics, and other safety-critical systems. The company's research is published through academic papers, textbooks, and tools, and it maintains a client login portal. The pages do not specify the product's packaging, such as free plans, trials, or editions.