Fork the consciousness, or download the project and create your own.

Benjamin Rosman

School of Computer Science and Applied Mathematics, University of the Witwatersrand

Builder The claim is about what the hardware does, and the philosophy is left open

Benjamin Rosman is a Professor in the School of Computer Science and Applied Mathematics at the University of the Witwatersrand in Johannesburg and director of the MIND Institute. His primary research program investigates how autonomous agents learn, compose, and transfer skills across complex decision spaces without hand-tuned reward shaping.

In consciousness-relevant agent modeling, Rosman collaborated with St John Grimbly, Jonathan Shock, Ryan Smith, and Mark Solms on the October 2026 study “Two Routes to Interoceptive Prioritisation: Preference Reweighting and Likelihood Precision in Multi-Need Homeostatic Agents” (arXiv:2610.18617). That work tests how an active-inference agent with a fixed total interoceptive precision budget can route attention toward whichever physiological channel is closest to a lethal boundary while dynamically scaling the urgency of its prior preferences.

His work brings formal sequential decision-making discipline to biologically inspired models of drive and affect, testing whether proposed homeostatic mechanisms actually keep an embodied agent alive under coupled action conflicts.

Known for. Reinforcement learning and decision-making in multi-need agents, skill composition, active inference homeostatic prioritisation

Coverage on this site

Related researchers