We introduce a new approach to unsupervised estimation of feature-rich semantic role la-beling models. Our model consists of two components: (1) an encoding component: a semantic role labeling model which predicts roles given a rich set of syntactic and lexi-cal features; (2) a reconstruction component: a tensor factorization model which relies on roles to predict argument fillers. When the components are estimated jointly to minimize errors in argument reconstruction, the induced roles largely correspond to roles defined in an-notated resources. Our method performs on par with most accurate role induction methods on English and German, even though, unlike these previous approaches, we do not incorpo-rate any prior linguistic knowledge about the languages. 1

Publisher's PDF, also known as Version of record

We present a new approach for unsuper-vised semantic role labeling that lever-ages distributed representations. We in-duce embeddings to represent a predi-cate, its arguments and their complex in-terdependence. Argument embeddings are learned from surrounding contexts involv-ing the predicate and neighboring argu-ments, while predicate embeddings are learned from argument contexts. The in-duced representations are clustered into roles using a linear programming formu-lation of hierarchical clustering, where we can model task-specific knowledge. Experiments show improved performance over previous unsupervised semantic role labeling approaches and other distributed word representation models. 1

We introduce the task of incremental se-mantic role labeling (iSRL), in which se-mantic roles are assigned to incomplete input (sentence prefixes). iSRL is the semantic equivalent of incremental pars-ing, and is useful for language model-ing, sentence completion, machine trans-lation, and psycholinguistic modeling. We propose an iSRL system that combines an incremental TAG parser with a seman-tically enriched lexicon, a role propaga-tion algorithm, and a cascade of classi-fiers. Our approach achieves an SRL F-score of 78.38 % on the standard CoNLL 2009 dataset. It substantially outper-forms a strong baseline that combines gold-standard syntactic dependencies with heuristic role assignment, as well as a baseline based on Nivre’s incremental de-pendency parser. 1

Inferring the information structure of scien-tific documents is useful for many NLP appli-cations. Existing approaches to this task re-quire substantial human effort. We propose a framework for constraint learning that re-duces human involvement considerably. Our model uses topic models to identify latent top-ics and their key linguistic features in input documents, induces constraints from this in-formation and maps sentences to their domi-nant information structure categories through a constrained unsupervised model. When the induced constraints are combined with a fully unsupervised model, the resulting model challenges existing lightly supervised feature-based models as well as unsupervised mod-els that use manually constructed declarative knowledge. Our results demonstrate that use-ful declarative knowledge can be learned from data with very limited human involvement. 1

cancer. He was 84 years old. Fillmore was one of the world’s pre-eminent scholars of lexical meaning and its relationship with context, grammar, corpora, and computation, and his work had an enormous impact on computational linguistics. His early theoret-ical work in the 1960s, 1970s, and 1980s on case grammar and then frame semantics significantly influenced computational linguistics, AI, and knowledge representation. More recent work in the last two decades on FrameNet, a computational lexicon and annotated corpus, influenced corpus linguistics and computational lexicography, and led to modern natural language understanding tasks like semantic role labeling. Fillmore was born and raised in St. Paul, Minnesota, and studied linguistics at the University of Minnesota. As an undergraduate he worked on a pre-computational Latin corpus linguistics project, alphabetizing index cards and building concordances. During his service in the Army in the early 1950s he was stationed for three years in Japan. After his service he became the first US soldier to be discharged locally in Japan, and stayed for three years studying Japanese. He supported himself by teaching English, pioneering a way to make ends meet that afterwards became popular with generations of young Americans abroad. In 1957 he moved back to the United States to attend graduate school at the University of Michigan. At Michigan, Fillmore worked on phonetics, phonology, and syntax, first in the American Structuralist tradition of developing what were called “discovery proce-dures ” for linguistic analysis, algorithms for inducing phones or parts of speech. Dis-covery procedures were thought of as a methodological tool, a formal procedure that linguists could apply to data to discover linguistic structure, for example inducing parts of speech from the slots in “sentence frames ” informed by the distribution of surround-ing words. Like many linguistic graduate students of the period, he also worked partly on machine translation, and was interviewed at the time by Yehoshua Bar-Hillel, who was touring US machine translation laboratories in preparation for his famous report on the state of MT (Bar-Hillel 1960).

is retained by the author(s) and / or other copyright owners and it is a condition of accessing these publications that users recognise and abide by the legal requirements associated with these rights. Take down policy The University of Edinburgh has made every reasonable effort to ensure that Edinburgh Research Explorer content complies with UK legislation. If you believe that the public display of this file breaches copyright please contact openaccess@ed.ac.uk providing details, and we will remove access to the work immediately and investigate your claim.