Bot to the future: using machine learning to develop the ultimate MW paper

Lightning Talk

Louise Rawlinson, Cogapp, United Kingdom, Gavin Mallory, Cogapp, UK, Tristan Roddis, Cogapp, UK

What happens when you data-mine twenty-one years of Museums and the Web conference papers, and then train a bot on the contents?

In this lightning talk we aim to show you the journey we took to create the ultimate, zeitgeist-capturing paper title that would wow reviewers in 2020. Marvel at how the crude beginnings of machine understanding morph into a sophisticated prediction engine, leveraging every buzzword-worthy machine-learning technique that we can think of along the way.

There should also be time for a bit of introspection: SGML and the information superhighway have fallen by the wayside; chatbots and machine-learning (oh no that’s us) are ascendant, but what are the perennial, enduring themes of the last twenty years?

Bibliography:
Natural Language Toolkit: https://www.nltk.org/
TF-IDF: https://en.wikipedia.org/wiki/Tf%E2%80%93idf
Markov chains: https://brilliant.org/wiki/markov-chains/