pySBD: Hidden Gem for Sentence Boundary Detection

Although it may seem simple, human language is noisy and complex. Only up to a certain point does dividing text into sentences based only on punctuation make sense. The best thing about pySBD is that it can handle a wide range of edge cases, including abbreviations, decimal numbers, and other challenging situations that are frequently seen in corpora from the legal, financial, and biomedical fields. PySBD recognises sentence boundaries using a rule-based method, in contrast to the majority of other […]

Read more

I used GPT-3 to answer Whatsapps

Before I talk about how the experiment went and how many friends I lost, it’s relevant to know what GPT-3 is. This ‘strangely’ named AI is just the acronym for ‘Generative Pre-trained Transformer 3’ and is a product from a company called OpenAI. It’s based around a relatively new machine learning model called a Transformer, this architecture has changed the game for sequential data models and has smashed its predecessor the RNN.

Read more
1 16 17 18 19 20 27