First of all my disclaimer: this is not an article for data scientists. They should already recognize the mistakes I formulated in this post title. Topic detection and sentiment analysis are classification problems in Natural Language Processing. BERT is an algorithm to solve, among the others, classification problems, but of course, I formulated a title helping business people understanding the topic 😉 Don’t expect a nerd discussion around BERT.


It is BERT a killer NLP application for business? Let’s say, considering an academic point of view is definitely a big step forward in NLP and Natural language Understanding. It is relatively young and we need to apply it to real business problems in order to better understand its impact. My experience is that when I work with academic data sets and benchmark my work with academic papers, you see definite improvements. When I work in the real business world, those improvements are less evident and the fine-tuning process is long and complex. My personal feeling is BERT can really bring benefits to NLP and NLU. I will keep you updated on how BERT helps sandsiv+ and, at the end, your business.

Sentiment analysis
Federico Cesconi

Read the article on LinkedIn

Start growing with sandsiv+ today