Abstractions and Generalization
Tim and Keith discuss the nature of abstractions in machine learning, questioning whether neural networks truly learn meaningful abstractions or just shortcuts. They delve into the importance of creating abstractions that are more generalizable out of distribution, emphasizing the need for machine learning to evolve beyond brittle, spurious abstractions.In this clip
From this podcast
Machine Learning Street Talk (MLST)
#063 - Prof. YOSHUA BENGIO - GFlowNets, Consciousness & Causality
Related Questions