Paris NLP Season 6 Meetup #4

Slim Frikha, ContentSquare

Unsupervised clustering of website URLs and clusters pattern recognition

At Contentsquare, we collect and analyze user interactions on the websites of our customers. To be able to produce insightful tips and recommendations, we combine this data with the website contextual data created thanks to a human manual configuration step.

An example of website contextualization is to know the web page category (product page, checkout, etc). Further, we would like to know what other web pages are similar to it within the website (we call them page groups).
In this talk, we will present a system that automates such website contextualization tasks using web pages URLs by: 1. creating business meaningful page groups 2. characterizing each page group with URL regexes 3. estimating the page category of each page group.

Mohamed Bamouh, Besedo

How can AI be used to summarize the text data?

Text Summarization is rewriting a consequently large piece of text into a smaller, shorter version while achieving minimal data loss and maximal information retention. In the NLP field, this task can be done by using specific algorithms and Deep Learning models, which take text data as input and either generate or extract a summary as output.

Our presentation on automatic text summarization will cover the most essential points, such as the types of Automatic Text Summarization, the evaluation metrics used to quantify the quality of a summary and we will present some state-of-the-art models. Also, we will show you a small application of text summarization in the form of a Twitter bot (homemade 😉 ) which summarizes abstracts of scientific papers in one sentence.

*** FIRST COME FIRST SERVED ! There will be a limit of 80 people. Make sure to arrive on time ;) See you soon ! ***

***