A Brief History of NLP leading to LLMs — Part 1
“If I have seen further than others, it is by standing upon the shoulders of giants” — Sir Isaac Newton
LLMs were not invented in a vacuum and they certainly were not built overnight. Instead, it took decades of research especially in the areas of deep learning and specifically in natural language processing (NLP) to build something as remarkable as LLMs. In this blog, we highlight some of the key moments in the decades-long journey so that we all can appreciate the contributions made by the many and get an idea of what may lie ahead of us.
1950s — Birth of NLP and Linguistic Theories
It is not a trivial task to get computers to understand human languages. NLP research starts with the task of translating sentences in one language to another. This task is called Machine Translation (MT).
In 1945, Warren Weaver, a researcher at the Rockefeller Foundation, was the first to present an information theory-based computer algorithm for MT. This led to academic research in this area and the first public demonstration of a machine translation project called Georgetown-IBM experiment, where, in 1954, researchers used IBM 701 computers to automatically translate 60 Russian sentences into English.
