Active Inference Insights
Jonas discusses the limitations of representational power in machine learning, using the analogy of pixel resolution in mapping. He emphasizes the advantage of smart computation over raw data capacity, illustrating how deeper insights can be gained by reallocating resources effectively. Tim highlights the significance of active inference, contrasting it with traditional machine learning by focusing on situational computation and agency.In this clip
From this podcast
Machine Learning Street Talk (MLST)
Jonas Hübotter (ETH) - Test Time Inference
Related Questions