| Classic algorithms for predicting structured data (eg., graphs, trees, etc.) rely on expensive (sometimes intractable) inference at test time. In this talk, I'll discuss several recent approaches that enable computationally efficient (eg., linear-time) prediction at test time. These approaches fall in the category of learning algorithms that optimize accuracy for some fixed notion of efficiency. I'll conclude by considering the question: can a learning algorithm figure out how to make fast predictions on its own?
|
