Today’s AI techniques have shown remarkable performance and are changing the world in which we live. However, the deep learning techniques that are leading to this AI revolution simply do not work with spherical data.

Kagenova has recently announced the fourpiAI platform for AI for spherical data.

fourpiAI unlocks the remarkable potential of deep learning for problems involving spherical data, in virtual reality and beyond.

fourpiAI dramatically outperforms standard approaches to apply AI to spherical data.

Get in touch about fourpiAI

Features

In fourpiAI, Kagenova has built efficient spherical AI techniques from the ground up so that they live natively on the sphere. This retains the symmetries of the sphere and eliminates any distortions due to projections. In addition, fourpiAI networks can be invariant to rotations of the original spherical data, which can significantly enhance performance. They are also scalable.

Preserve symmetries

Eliminate distortions

Rotational equivariance

Scalable