2017/02/19 by Ann Irvine, Mark Dredze, Irvine, Ann +1
Computer Science · #Computation and Language (cs.CL) #FOS: Computer and information sciences #Natural Language Processing Techniques #Speech Recognition and Synthesis #Speech and dialogue systems
paper · pdf · doi:10.48550/arxiv.1702.05793
openalex publication_date 2017/02/19 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
This work presents a systematic theoretical and empirical comparison of the\nmajor algorithms that have been proposed for learning Harmonic and Optimality\nTheory grammars (HG and OT, respectively). By comparing learning algorithms, we\nare also able to compare the closely related OT and HG frameworks themselves.\nExperimental results show that the additional expressivity of the HG framework\nover OT affords performance gains in the task of predicting the surface word\norder of Czech sentences. We compare the perceptron with the classic Gradual\nLearning Algorithm (GLA), which learns OT grammars, as well as the popular\nMaximum Entropy model. In addition to showing that the perceptron is\ntheoretically appealing, our work shows that the performance of the HG model it\nlearns approaches that of the upper bound in prediction accuracy on a held out\ntest set and that it is capable of accurately modeling observed variation.\n