Non conceptual knowledge systems

This lecture explores the recent successes of Deep learning and its impact on Digital Humanities, focusing on the encoding/decoding of writing systems and images. It delves into the emergence of non conceptual knowledge, its characteristics, and its application in articulated reasoning. The lecture discusses the challenges of grounding real-world data into logical symbols and the potential of non conceptual knowledge in performing complex tasks. It also highlights the recent advancements in non conceptual knowledge systems, such as Neural Machine Translation and Unsupervised NLP. The lecture concludes by examining the limitations of Recurrent Neural Networks, the introduction of Transformers, and the revolutionary capabilities of GPT-3 in the attention economy and few-shot learning.