Pioneering AI through Linguistics: 5 Linguists who Ushered AI in Language
As Artificial Intelligence continues to transform the way we engage with languages, the field of linguistics has played a vital role in developing the language abilities of AI by aiding in processes such as machine learning, speech recognition, and natural language processing. Let us then learn about some of the pioneering linguists whose contributions to the study of language and analysis of the role of cognition in linguistics have directly or indirectly allowed AI to acquire language:
Noam Chomsky: Chomsky is the most preeminent linguist we have in our world today. Chomsky’s most important contribution to linguistics has been his theory of language acquisition which argues that the structure of the human brain naturally allows humans to have the capacity to learn, acquire, and use languages. Thus, language acquisition is intrinsic in humans. Chomsky’s work on generative grammar, which identifies the underlying common structures in seemingly disconnected sentences, has instrumentally shaped how natural language processing is deployed in AI. This is because Chomsky’s theory rejects the notion that every language is unique and instead argues how all languages assume a universal set of formal suppositions. Recently, Chomsky published an opinion piece in the New York Times that argued against alarmists conceptions of the dangers of AI and claimed the superiority of the human mind over AI.
Karen Sparck Jones: Jones is a foundational figure in the realm of natural language processing. Her theories have been of immense use in teaching computers to understand human language and have formed the foundation of search engines like Google. Her influential paper in 1972 used a combination of statistics and linguistics to establish formulae that enabled computers to gauge the relationship between different words.
Aravid Joshi: Joshi was the Professor of Computer and Cognitive Science at the University of Pennsylvania. Joshi’s definition of the tree-adjoining grammar (TAG), that borrows from linguistics, psychology, and computer science, has deeply enhanced our understanding of computational linguistics. His detailing of the mildly sensitive context-sensitive TAG accounted for the syntax of human languages while at the same time being amenable to being tracked by computers.
Roger C Schank: Schank used a combination of linguistics, computing, and cognitive sciences to delve deep into the nature of the human mind and its language abilities to foster building computer models based on this understanding. He formulated the “conceptual dependency theory” which would allow AI to represent its knowledge in a way that would resemble natural language by making meaning independent of words such that two sentences that mean the same regardless of the words that compose them would have a single representation. Schank also worked towards developing ways to assemble raw materials of knowledge called “scripts'' in a way that would allow computers to use this information in a way similar to how humans use past experiences. He termed this “case-based reasoning”.
Terry Winograd: Winograd has contributed significantly to the field of AI. His research focuses on natural language understanding by computers and enhancing human-computer interaction. He conceptualized the SHRDLU program which would allow computers to understand natural language through simple dialogues about a small world of objects (called the Blocks World). The program mimics a robot who understands commands in English, performs tasks, and answers questions pertaining to the Blocks World.
Sources:
https://www.nytimes.com/2023/03/08/opinion/noam-chomsky-chatgpt-ai.html
https://www.nytimes.com/2019/01/02/obituaries/karen-sparck-jones-overlooked.html
https://direct.mit.edu/coli/article/44/3/387/1597/Aravind-K-Joshi
https://www.nytimes.com/2023/02/20/technology/roger-c-schank-dead.html
https://research-api.cbs.dk/ws/portalfiles/portal/58908562/6460.pdf