[Pycon] [new paper] "Harshdeep Singh" - Text Summarisation made fun!

[info a pycon.it]

Title: Text Summarisation made fun!
Duration: 60 (includes Q&A)
Q&A Session: 15
Language: en
Type: Talk

Abstract: The goal of this talk is to explain this quote : “You shall know a ‘word’ by the company it keeps!”. In this talk, we will go through as to how to build a model for text summarisation and its possible applications in the real world scenario. The prospective audience should have a basic understanding of neural networks and natural language processing.

We will provide an intuitive explanation (the talk would not be all mathematical!) as to how to do the data preprocessing for a large dataset. We will also talk about how certain Python libraries make it easier to structure a machine learning pipeline.

For example: Given an article from newspap, the model will generate a summary of 100 words.

We will also walk through the various caveats while building these kinds of complex models and how to circumvent these. The talk will also include the possible way to evaluate and test the model to ensure it is working as expected.

Tags: [u'Machine Learning', u'Python', u'tensorflow', u'text-analysis', u'natural-language-processing', u'numpy', u'summarisation']