10:30 AM - 11:30 AM
Hosted by SMWHH
Markthalle - Marx Klosterwall 11 Hamburg Hamburg 20095
For Social Media Analysis, more and more natural language processing (NLP) methods are adopted to get a handle at content analysis: While you can get a long way with counting likes and re-tweets, you need some amount of language understanding to let your analysis software see what’s inside a text.
In research, there has been a steady shift away from rule-based modeling towards statistical learning. In industrial applications, however, list-based approaches still dominate: sentiment analysis, for instance, is implemented by counting the number of positive and negative words in a sentiment lexicon.
Whether fancy machine learning or matching word lists, natural language components used in today’s NLP pipelines are static in the sense that they are created once, subsequently applied without further change. Even if a component performs well today, it might already be outdated in a few months from now, as trends in Social Media have shifted.
In this talk, I will motivate an adaptive approach to natural language processing, where NLP components get smarter through usage over time through continuous feedback. With the help of recent research prototypes, three stages of data-driven adaptation will be illustrated: feature/resource induction, induction of processing components and continuous data-driven learning. These give rise to NLP technology that adapts to the task at hand, reaching much higher levels of performance than dictionary-based approaches.
*Event language is english * Die Veranstaltung ist englischsprachig*
*Important: times and locations are subject to change/ Wichtig: Zeiten und Orte können sich kurzfristig ändern*
Venue & Location
Markthalle - MarxKlosterwall 11
Hamburg, Hamburg 20095