Textbook of AI

1946–, Computer scientist

Terry Winograd is an American computer scientist whose 1970 MIT PhD thesis produced SHRDLU, a natural-language understanding system that conversed about a "blocks world" of coloured blocks and pyramids on a virtual table. SHRDLU could parse complex sentences ("Find a block which is taller than the one you are holding and put it into the box"), execute the implied actions in its simulated world, and answer follow-up questions including those requiring knowledge of its own past actions ("Why did you do that?"). The system was the most impressive demonstration of natural-language understanding of its decade.

Winograd became disenchanted with the AI programme during the 1970s and reoriented his career towards human–computer interaction. His 1986 book with Fernando Flores, Understanding Computers and Cognition, argued that the symbolic-AI tradition, which he had himself contributed to, fundamentally mischaracterised human cognition. He moved to Stanford, where he was a founding faculty member of the Stanford Human–Computer Interaction Group and supervised the doctoral work of Larry Page on what would become Google.

The Winograd Schema Challenge, proposed by Hector Levesque in 2011 and named in honour of Winograd's early ambiguity examples, was for several years a leading benchmark for commonsense reasoning in NLP. Modern large language models now solve it at near-human level.