HomeResearchGroups

The logo of BONXAI-lab. A bonsai tree shaped like a brain and with the text BONXAI
The logo of BONXAI-labPhoto: Image generated by DALL-E 2 (OpenAI)

BONXAI-lab is a research lab for Biologically inspired self-Organizing Networks for eXplainable and sustainable Artificial Intelligence.

We build upon classical statistics and computational neuroscience to develop novel algorithms for artificial intelligence and machine learning. Our work takes inspiration from biology and is exploiting statistical knowledge on uncertainty, parsimony and stochastic processes to build more sustainable AI.

The lab is part of the Bioinformatics and Applied Statistics (BIAS) group at KBM at the Norwegian University of Life Sciences.

About

  • The activity at BONXAI-lab is part of a new national initiative for NeuroAI aiming to involve several renowned national research environments in the intersection between Neuroscience and AI.

    A central focus area of BONXAI-lab is to explore neural nets as dynamic Complex Adaptive Systems (CAS), emerging partly in a self-organized manner from simple learning rules and simple regulation mechanisms. This approach, where complexity is grown from simplicity, is fundamentally different from the bottom-up engineering approach which is dominating the field of AI today.

    By applying classical statistical methods with Predictive Coding Theory and Bayesian learning as frames of reference, we anticipate that knowledge from this project will not only contribute to a more sustainable path for AI, but also add to our understanding of human cognition and learning in general.

  • We collaborate with several research environments as member of NORA – Norwegian Artificial Intelligence Research Consortium, and as part of the NeuroAI initiative in Norway.

Members

Publications

TBC