Computers that understand human texts in such a way that they can fully comprehend the meaning and completely rewrite them fully autonomously: recent developments show that this is no longer a matter of fantasy, but of reality.
Natural Language Processing (NLP) refers to an expertise dedicated entirely to the interaction between computers and linguistics. Thanks to Data Science and AI, NLP algorithms can analyze texts to extract information, which can then be used by the algorithm to generate its own complete texts.
NLP as a field of expertise is nothing new and has been around for over 60 years. Especially in the last few decades, developments have been very fast thanks to the use of neural networks and Big Data. Furthermore, Deep Learning has become the basis of innovations from a plethora of AI-related fields.
But the very latest innovation is taking place in the form of the GPT-3 model (Generative Pre-trained Transformer), a powerful model containing over 175 billion parameters, developed by OpenAI. The GPT-3 model brings along a lot of new, exciting and innovative capabilities. For example, it can generate coherent texts in response to a human political argument without a single example. This phenomenon is labeled as “zero-shot training” and is a testament to the improved linguistic understanding of GPT-3.
In short, GPT-3 opens a lot of new doors, also for Squadra Machine Learning Company. Are you curious about all the possibilities of GPT-3? Read the full article on Emerce.
(Please note that the article is unfortunately only available in Dutch)