A Brief History of Artificial Intelligence
Ron Fulbright · 2020
Dating back to even before the founding of the field of artificial intelligence in 1955, visionaries imagined machines able to think and learn on their own. Research in artificial intelligence has enjoyed an active 75-year history. This chapter discusses milestone achievements in artificially intelligent computer programs, machine learning, game playing, and chatbots. The chapter also describes the two main approaches to knowledge representation developed over the decades: symbolic processing (rule processing) and connectionism (neural networks). Recently, multi-layer neural networks have achieved super-human performance using unstructured training data and unsupervised learning techniques. Cognitive systems will be able to acquire expert-level knowledge and ability on their own. A key area of research throughout the history of artificial intelligence has been natural language understanding and speech with significant advances having been made in recent years. People are now accustomed to talking to their smartphones and voice-controlled assistants. This heralds the very beginning of the cognitive systems era in which people will routinely carry on conversations with cognitive systems using natural language just as if talking to another person. Interacting with cognitive systems in this manner insures the infusion of cognitive system technology into the very fabric of our lives.