Anna Rogers (University of Massachussetts Lowell, USA)
The proposed suite of two courses introduce the foundations of NLP, an area rapidly growing both in terms of academic developments and the number of jobs. First, the introductory course will cover corpus pre-processing pipelines, basic machine learning experiments on text classification, and analysis of their results. Second, the advanced course will introduce the basic classes of deep learning models and discuss recent trends in neural net-based distributional representations at the word, subword and sentence levels. These courses are aimed at linguists with only basic programming skills in Python, and will highlight in particular the problems with the current evaluation paradigms, dataset design, and general methodological challenges — the interdisciplinary areas in which their expertise is much needed.