ABSTRACT
The study of Artificial Intelligence (AI) has been pursued from the very beginning in two different styles, jokingly referred to as scruffy and neat. These styles actually reflect distinct viewpoints of the discipline and its objectives. In this paper, we review the tension between scruffies and neats over the history of AI. We analyze the impact of current deep learning methods in this debate, suggesting that the development of broad computational architectures is a particularly promising path for AI.
KEYWORDS:
Artificial Intelligence; Logic; Knowledge representation; Deep learning