How do words evoke meaning?
Can we model this process statistically?
How do we estimate these types of models from large amounts of incomplete data, balancing computational and statistical efficiency?
machine learning, natural language processing.
natural language semantics,
statistical learning theory,
method of moments,
SEMPRE: toolkit for building semantic parsers with execution ("programming by natural language")
CodaLab: platform for reproducible research ("Git for computational experiments")
MLcomp: website for comparing machine learning algorithms on datasets (with Jacob Abernethy)
Braque: service that notifies you of new research papers (with Hal Daume)