The most popular and comprehensive Open Source ECM platform
GPT-3 is the super Natural Language Processing (NLP) algorithm from OpenAI that was trained on nearly all text (hundreds of billions of words) from the internet. Given some prompts, it is able to generate essays and text that are amazingly good and would be mistaken as something written by a human.
The GPT-3 algorithm is all about identifying patterns in language. From the patterns it identifies, it assigns parameters. GPT-3 has 175 billion different parameters which can be tuned to generate the text that it spews out.
It turns out that the GPT-3 algorithm works pretty well on identifying patterns in nonstandard text too, like computer code and spreadsheets.
It turns out that NLP algorithms like GPT-3 are able to help scientists in studying how viruses like COVID mutate.
Researchers at MIT said that “we were excited about recent advances in NLP language models for understanding human language by training them on raw text alone. We thought that since the most abundant data for viruses is just raw viral sequence, we could also learn very complex patterns from viral sequence datasets by training a language model.”
It turns out that there are striking similarities between English grammar and syntax and amino acid ordering in proteins. Using NLP, the researchers were able to predict potential mutations of COVID and other viruses.