Learning Challenges
The conversation highlights the struggle between relying on machines to infer knowledge from data versus the necessity of human-engineered specifications. While one perspective advocates for developing unsupervised learning algorithms, another emphasizes the importance of active learning, where human guidance shapes the machine's learning process. The complexities of encoding knowledge into systems remain a significant hurdle in advancing robust AI.In this clip
From this podcast
Lex Fridman Podcast
Gary Marcus: Toward a Hybrid of Deep Learning and Symbolic AI | Lex Fridman Podcast #43
Related Questions