Transformative Advances in NLP and AI: Charting the Evolution of Technology

This paper attempts to enter the world of NLP (human language processing) from the three perspectives of physics, mathematics and computer science.The article explains why science has chosen word-vectors and word vectorization in NLP and describes the 2 models that have established themselves in this world of words: Word2Vec and GloVe.After having a clear picture of how artificial intelligence deals with words and human language processing, the topics of Time and Attention are treated in the new approach of Google which has already moved to another paradigm in word processing: the BERT models, transformers and Attention mechanisms.This answers the questions why Time and temporal recurrence have been abandoned in favor of models with transformers and Attention mechanisms.The paper also includes an explanation of the complex processes that take place inside the Transformer for a simple translation from one language to another.