2026-09-21T01:03:50Zhttps://api.repository.cam.ac.uk/server/oai/request

oai:www.repository.cam.ac.uk:1810/2700262024-06-26T13:52:11Zcom_1810_219481com_1810_256065col_1810_219482

Structured Learning with Inexact Search: Advances in Shift-Reduce CCG Parsing 10.17863/CAM.16866 Xu, Wenduan Clark, Stephen Statistical shift-reduce parsing involves the interplay of representation learning, structured learning, and inexact search. This dissertation considers approaches that tightly integrate these three elements and explores three novel models for shift-reduce CCG parsing. First, I develop a dependency model, in which the selection of shift-reduce action sequences producing a dependency structure is treated as a hidden variable; the key components of the model are a dependency oracle and a learning algorithm that integrates the dependency oracle, the structured perceptron, and beam search. Second, I present expected F-measure training and show how to derive a globally normalized RNN model, in which beam search is naturally incorporated and used in conjunction with the objective to learn shift-reduce action sequences optimized for the final evaluation metric. Finally, I describe an LSTM model that is able to construct parser state representations incrementally by following the shift-reduce syntactic derivation process; I show expected F-measure training, which is agnostic to the underlying neural network, can be applied in this setting to obtain globally normalized greedy and beam-search LSTM shift-reduce parsers. University of Cambridge 2017-12-01 Thesis Doctoral Doctor of Philosophy (PhD) en The Carnegie Trust for the Universities of Scotland; The Cambridge Trust https://www.repository.cam.ac.uk/handle/1810/270026 https://apollo8-f-pro.lib.cam.ac.uk/bitstreams/30ff7cd1-1e3f-4b94-9d00-d31621b5f7e2/download 87eda9de84448d1f82354d60eee3eb5f https://apollo8-f-pro.lib.cam.ac.uk/bitstreams/211573f1-6264-42ba-889a-da11beaf12e9/download a7dffa29b57e5c4a8983caa12e16be2c https://www.rioxx.net/licenses/all-rights-reserved/ Structured Prediction Structured Learning with Inexact Search Violation-Fixing Structured Perceptron Recurrent Neural Networks LSTMs Combinatory Categorial Grammar Shift-Reduce Transition-based Parsing