HELLO WORLD: The Inexplicable Efficacy of Language Models

摘要

The goal of teaching language to machines has existed since the early days of computing, perhaps since the Turing Machine was conceived in 1936. This pursuit was inspired by and has further influenced applications like audio transcription, translation of books, information retrieval, summarizing complex writings, etc. This column provides a brief insight into language modeling-a statistical way of teaching language to machines, and its inexplicable capabilities. Furthermore, it highlights a few shortcomings of such models and discusses some recent approaches to address them.