AI, Machine Learning & Media

2016.12.05

Date Fri 16 Dec
Time 12:00 13:30
Location AU DataLab (CAVI), Aabogade 34D bldg. 5345 8200 Aarhus N

Datalab has received funding for a research project on AI in Digital Sociology and kick-starts this by inviting two interesting speakers to address the theme of AI, Machine Learning and Media. Both speakers focus on processing large amount of text through machine learning algorithms and will talk about the potentials and challenges in doing so. Business partner from Digital Society Mads Rydahl is Innovation Executive & Co-founder at the company Unsilo. He has lived 5 years in Silicon Valley, worked at Stanford University, and was head of Product and Design at Siri.com. He has built games for Lego Mindstorms, interfaces for Bang & Olufsen, authored a portfolio of patents acquired by Apple, and created the world’s best casual game. Leon Derczynski from AU Datalab is from computational linguistics (Natural Language Processing) and has been researching and publishing extensively on the processing of social media Twitter data with the use of machine learning. He is associate professor at Sheffield University and visiting researcher at UC San Diego in the spring of 2017.

 

Artificial Intelligence for research

UNSILO works with leading Scientific Publishers to enrich their content and improve discoverability across domains and disciplines. Our discovery tools capture trending ideas and novel concepts as they emerge, and they help researchers find articles that describe parallel research of similar ideas across different domains and disciplines. In this talk I will present our vision, describe some of the major challenges we are currently trying to solve, and outline the future direction of what we call Text Intelligence.

 

Why AI needs the crowd and your social media

We're in the wild west times of a burgeoning new era of artificial intelligence. One of the greatest challenges for AI is to talk with us and understand the reply - and indeed this is the current sign for intelligence (via the Turing test). For years, we've built and learned from large datasets of formal language, often news, in order to try to understand how the process works. But the strikingly different reality is that not only do humans not communicate like this - it also gives our systems strong gender, socio-economic and racial biases, that leave them ill-equipped to understand us. Based on real way we use language, this talk shows how crippled existing text processing systems really are, and that the experts we trust to improve them are really incapable of doing so - for that, we must have the crowd.

Time and Place: 16 December 2016 at 12.00-13.30 followed by sushi, snacks and cava reception for Datalab and Digital Society, in AU Datalab – see directions here Register by sending email no later than 12 December to Research Group Coordinator: simonerosengreen@cc.au.dk . Please indicate if you are attending talks and/or reception.

Event type