Modeling Prediction in Human Parsing
Speaker: Frank Keller
Institution: University of Edinburgh
Abstract:
There is strong evidence that human sentence processing is incremental, i.e., that structures are built word by word. Recent experiments show that the processor also predicts upcoming linguistic material on the basis of previous input. We present a computational model of human parsing that is based on a variant of tree-adjoining grammar and includes an explicit mechanism for generating and verifying predictions, while respecting incrementality and connectedness. We outline a linking function that explains processing difficult as a combination of prefix probability (surprisal) and verification cost. The resulting model is able to capture both locality effects and surprisal effects, unlike existing models of human parsing. We evaluate the model against reading time data from relative clause and coordinate structure experiments.
Joint work with Vera Demberg.