22 January 2020
The Best NLP Tools of Early 2020: Live Demos
Igor Kaufman, Principal Consultant at DataArt, shares his list of the cutting- edge Natural Language Processing (NLP) and Natural Language Understanding (NLU) tools.
«2019 was the year of NLP. The cutting edge models developed by Google, OpenAI, Facebook, and others became publicly available for a wider audience.»
«BERT — Understanding Texts. BERT is a pre-trained model published by Google and is intended to better understand what people search for.»
«GPT-2 — Creating Texts. Its main purpose is to predict the next word, given all of the previous words within a text.»
«SpaCy — Implementing NLP in Production. SpaCy is a free open-source NLP library developed by ExplosionAI. It’s aimed at helping developers in production tasks.»
«AllenNLP — A Famous Alternative. Textual Entailment takes a pair of sentences and predicts whether the facts in the first necessarily imply the facts in the second one. It doesn’t always work well, but that’s one of the challenges of conversational AI to reveal these kinds of implications.»
«Text Summarization — TL;DR. This demo uses the extraction-based approach: the only thing you need to do is to tell the number of sentences you want to see as a result. The abstraction-based approach lets the machine rephrase the text in a shorter way.»
«Google AutoML Natural Language. In addition to what Google AutoML can do, in IBM Watson you may check the emotional characteristics and the concepts described in the text.»
«HuggingFace — Very Useful for Production. HuggingFace makes different NLP models easy to use in production by training them additionally or wrapping them into easily pluggable libraries.»
«Berkley Neural Parser. The Berkley parser annotates a sentence with its syntactic structure by decomposing it into nested sub-phrases.»
Original article can be found here.