The Learning of the Past Tense of Danish Verbs: Language Learning in Neural Networks

1994 ◽  
Vol 15 (1) ◽  
pp. 15-35 ◽  
Author(s):  
K. A. JENSEN ◽  
I. ULBAeK
Keyword(s):  
Neural Networks ◽  
Language Learning ◽  
Past Tense ◽  
The Past

scholarly journals Recurrent Neural Networks in Linguistic Theory: Revisiting Pinker and Prince (1988) and the Past Tense Debate

2018 ◽  
Vol 6 ◽  
pp. 651-665 ◽  
Author(s):  
Christo Kirov ◽  
Ryan Cotterell
Keyword(s):  
Neural Networks ◽  
Cognitive Modeling ◽  
Past Tense ◽  
Network Architectures ◽  
The Past ◽  
Current State ◽  
Empirical Performance ◽  
English Verb

Can advances in NLP help advance cognitive modeling? We examine the role of artificial neural networks, the current state of the art in many common NLP tasks, by returning to a classic case study. In 1986, Rumelhart and McClelland famously introduced a neural architecture that learned to transduce English verb stems to their past tense forms. Shortly thereafter in 1988, Pinker and Prince presented a comprehensive rebuttal of many of Rumelhart and McClelland’s claims. Much of the force of their attack centered on the empirical inadequacy of the Rumelhart and McClelland model. Today, however, that model is severely outmoded. We show that the Encoder-Decoder network architectures used in modern NLP systems obviate most of Pinker and Prince’s criticisms without requiring any simplification of the past tense mapping problem. We suggest that the empirical performance of modern networks warrants a reexamination of their utility in linguistic and cognitive modeling.

Preverbation and narrativity in Lithuanian. The distribution of finite simplex and compound verbs in narrative main clauses

2010 ◽  
Vol 1 ◽  
pp. 167-192
Author(s):  
Lea Sawicki
Keyword(s):  
Narrative Text ◽  
Past Tense ◽  
The Past ◽  
Verb Forms ◽  
Compound Verbs ◽  
Finite Verb

The article deals with the use of simplex and compound (prefixed) verbs in narrative text. Main clauses comprising finite verb forms in the past and in the past habitual tense are examined in an attempt to establish to what extent simplex and compound verbs exhibit aspect oppositions, and whether a correlation exists between the occurrence of simplex vs. compound verbs and distinct textual units. The investigation shows that although simple and compound verbs in Lithuanian are not in direct aspect opposition to each other, in the background text portions most of the verbs are prefixless past tense forms or habitual forms, whereas in the plot-advancing text portions, the vast majority of verbs are compound verbs in the simple past tense.  

Learning Zulu

2016 ◽  
Author(s):  
Mark Sanders
Keyword(s):  
Language Learning ◽  
Cultural History ◽  
South African ◽  
Press Coverage ◽  
African History ◽  
African Culture ◽  
The Past ◽  
Correct Form ◽  
Linguistic Purism

When this book's author began studying Zulu, he was often questioned why he was learning it. This book places the author's endeavors within a wider context to uncover how, in the past 150 years of South African history, Zulu became a battleground for issues of property, possession, and deprivation. The book combines elements of analysis and memoir to explore a complex cultural history. Perceiving that colonial learners of Zulu saw themselves as repairing harm done to Africans by Europeans, the book reveals deeper motives at work in the development of Zulu-language learning—from the emergence of the pidgin Fanagalo among missionaries and traders in the nineteenth century to widespread efforts, in the twentieth and twenty-first centuries, to teach a correct form of Zulu. The book looks at the white appropriation of Zulu language, music, and dance in South African culture, and at the association of Zulu with a martial masculinity. In exploring how Zulu has come to represent what is most properly and powerfully African, the book examines differences in English- and Zulu-language press coverage of an important trial, as well as the role of linguistic purism in xenophobic violence in South Africa. Through one person's efforts to learn the Zulu language, the book explores how a language's history and politics influence all individuals in a multilingual society.

Beyond Deep Learning: An Econometric Example

2020 ◽  
Vol 28 (Supp01) ◽  
pp. 31-38 ◽  
Author(s):  
Ruofan Liao ◽  
Paravee Maneejuk ◽  
Songsak Sriboonchitta
Keyword(s):  
Neural Networks ◽  
Deep Learning ◽  
Prediction Models ◽  
Original Data ◽  
Parametric Model ◽  
Parametric Models ◽  
The Past ◽  
Currency Exchange ◽  
Currency Exchange Rate ◽  
Linear And Nonlinear

In the past, in many areas, the best prediction models were linear and nonlinear parametric models. In the last decade, in many application areas, deep learning has shown to lead to more accurate predictions than the parametric models. Deep learning-based predictions are reasonably accurate, but not perfect. How can we achieve better accuracy? To achieve this objective, we propose to combine neural networks with parametric model: namely, to train neural networks not on the original data, but on the differences between the actual data and the predictions of the parametric model. On the example of predicting currency exchange rate, we show that this idea indeed leads to more accurate predictions.

scholarly journals Laplacian networks: bounding indicator function smoothness for neural networks robustness

2021 ◽  
Vol 10 ◽  
Author(s):  
Carlos Lassance ◽  
Vincent Gripon ◽  
Antonio Ortega
Keyword(s):  
Neural Networks ◽  
Deep Learning ◽  
Supervised Learning ◽  
Indicator Function ◽  
Training Data ◽  
Theoretical Justification ◽  
The Past ◽  
Noisy Examples

For the past few years, deep learning (DL) robustness (i.e. the ability to maintain the same decision when inputs are subject to perturbations) has become a question of paramount importance, in particular in settings where misclassification can have dramatic consequences. To address this question, authors have proposed different approaches, such as adding regularizers or training using noisy examples. In this paper we introduce a regularizer based on the Laplacian of similarity graphs obtained from the representation of training data at each layer of the DL architecture. This regularizer penalizes large changes (across consecutive layers in the architecture) in the distance between examples of different classes, and as such enforces smooth variations of the class boundaries. We provide theoretical justification for this regularizer and demonstrate its effectiveness to improve robustness on classical supervised learning vision datasets for various types of perturbations. We also show it can be combined with existing methods to increase overall robustness.

Four decades of ELT development in Hong Kong: Impact of global theories on the changing curricula and textbooks

2019 ◽  
pp. 136216881986556
Author(s):  
Jim Yee Him Chan
Keyword(s):  
Hong Kong ◽  
Language Learning ◽  
Language Teaching ◽  
Future Research ◽  
Oxford University ◽  
The Past ◽  
Framework Approach ◽  
Language Needs ◽  
Recent Debate ◽  
Textbook Design

The past 40 years have witnessed significant developments in ELT research, reflecting the changes in learners’ language needs and the extensive development of various language learning/teaching methods in different times and places. The aim of this study is to provide a systematic and comprehensive account of changing ELT methods (oral-structural approach, communicative language teaching and task-based language teaching) in Hong Kong’s secondary education between 1975 and the present. By adopting Richards and Rodgers’s (2014) framework (approach, design and procedure), it examined how ELT theories have been transformed into local curricula (1975, 1983, 1999 and 2002/07) and commercial textbooks (Longman, Oxford University Press) via detailed content analysis. The findings suggest that research into ELT methods in Hong Kong over the past decades has generally directed the designs of the language curricula. Changes in the textbooks, however, have been relatively limited, although considerable attempts have been made to align textbook design with ELT trends. By considering various constraints in the theory-to-practice process, this study offers suggestions for future research and language teaching, particularly regarding the recent debate over the choice between the ‘weak’ and ‘strong’ versions of task-based language teaching in EFL contexts, and the post-methods perspective in language teaching.

SYSTEM OF PARTICIPLES IN THE OSSETIAN LANGUAGE COMPARED TO THE RUSSIAN LANGUAGE.

2020 ◽  
Author(s):  
З.И. Годизова ◽  
Д.В. Габисова
Keyword(s):  
The State ◽  
Russian Language ◽  
Past Tense ◽  
Language Family ◽  
European Language ◽  
Parts Of Speech ◽  
Imperfective Aspect ◽  
The Past ◽  
The Russian Language ◽  
Conversion Transition

Актуальность предпринятого исследования обусловлена тем, что причастие в современном осетинском языке не привлекало активного внимания ученых, имеются лишь общие описания причастий, а специальные исследования, посвященные причастиям, практически отсутствуют. Представляется интересным и актуальным сравнение системы причастий и их грамматических особенностей в осетинском и русском языках. Этот интерес обусловлен принадлежностью сопоставляемых языков к общей индоевропейской семье языков, а также тесным их взаимодействием в условиях двуязычия, что, очевидно, может отразиться и на системе причастий. Научная новизна данной статьи заключается в том, что в ней исследуются грамматические особенности всех разрядов причастий в осетинском языке в сопоставлении с русским языком. На основании проведенного анализа установлено, что в современном осетинском языке система причастий включает пять разрядов, разнообразных в своих грамматических проявлениях, в степени регулярности, в склонности переходить в состав других частей речи. Выявлены наиболее значительные отличия осетинских причастий от русских: существование причастий будущего времени в системе осетинского языка, отсутствие у причастий показателей времени и залога, а также именных грамматических категорий (падежа, числа, рода). Установлено также, что в осетинском языке категория вида в большей степени управляет категорией времени, в силу чего несовершенный вид причастий предполагает только настоящее время, а совершенный только прошедшее отсутствует четкая залоговая оппозиция причастий в осетинском языке. Определено также, что осетинские причастия не имеют членных (полных) форм, но функционируют в роли и сказуемого, и определения, хотя в большей степени тяготеют к предикативной роли. В осетинском языке причастия гораздо менее употребительны сравнительно с причастиями в русском языке и чаще вступают в отношения грамматической омонимии с другими частями речи. The relevance of the undertaken study is determined by the fact that participles in the modern Ossetian language are still insufficiently studied. There are only the most general descriptions of grammar features of participles. The comparison of the system of participles and their grammar features seems interesting and actual, especially considering the fact that the Ossetian and Russian languages belong to different groups of the Indo-European language family. Furthermore, in the context of bilingualism the Russian and Ossetian languages interact actively and that can affect the system of participles. The scientific novelty of the article is determined by the fact that it examines the grammatical features of all categories of participles in the Ossetian language in comparison with the Russian language. The conducted research allowed to elicit five categories in the system of participles in the modern Ossetian language. The analysis of the results showed the participles are diverse in their grammatical characteristics, in the degree of regularity, and in the tendency to transition into other parts of speech. The research defined the most significant differences between Ossetian and Russian participles: existence of future participles in the system of the Ossetian language absence of adjectival grammar categories of gender, number and case as well as formal markers of tense and voice in Ossetian participles. The tense category in Ossetian subordinates to the aspect category to a far greater extent therefore the imperfective aspect of participles accepts the present tense forms only, while perfective acts in the past tense forms Ossetian participles lack explicit voice opposition. Ossetian participles do not have full forms, but they can have syntactic functions of both the predicate and the attribute in a sentence, although the predicative function is more typical for them. Participles in the Ossetian language are much less common compared to participles in Russian and are more disposed to conversion (transition to the category of nouns, verbal adverbs, adjectives, words of the state category).

Sign in / Sign up

Export Citation Format

Share Document