Applied statistical methods developed in population genetics to linguistic time series to distinguish drift from selection in large text data sets.

Used mathematical and statistical methods to model the causes of language change in large historical text data sets, 8 publications, 20 presentations at national and international conferences.

Collaborated on international team to manually adjusted automatic alignment of broad and narrow transcriptions to speech signal to generate high-quality facial animations.

Worked on international team to evaluate the quality and accuracy of automatically generated summaries of search engine results, evaluated fit of advertisements to websites, edited automatically generated ad summaries, formulated heuristics to improve summarization methods.

Constructed portions of the Columbia Games Corpus. Transcribed audio from pairs of speakers playing an interactive game, manually adjusted automatic alignment of transcriptions to speech signal, labelled different classes of conversationalpatterns.