scholarly journals Zero-Shot Learning on Semantic Class Prototype Graph

2018 ◽  
Vol 40 (8) ◽  
pp. 2009-2022 ◽  
Author(s):  
Zhenyong Fu ◽  
Tao Xiang ◽  
Elyor Kodirov ◽  
Shaogang Gong
Keyword(s):  
Semantic Class

The Sound of Enemies and Friends in the Neighborhood

2011 ◽  
Vol 58 (6) ◽  
pp. 454-463 ◽  
Author(s):  
Diane Pecher ◽  
Inge Boot ◽  
Saskia van Dantzig ◽  
Carol J. Madden ◽  
David E. Huber ◽  
...  
Keyword(s):  
Word Recognition ◽  
Visual Word Recognition ◽  
Neighborhood Effects ◽  
Classification Performance ◽  
Visual Word ◽  
Orthographic Neighborhood ◽  
Semantic Classification ◽  
Semantic Class ◽  
Orthographic Neighbors ◽  
Target Words

Previous studies (e.g., Pecher, Zeelenberg, & Wagenmakers, 2005) found that semantic classification performance is better for target words with orthographic neighbors that are mostly from the same semantic class (e.g., living) compared to target words with orthographic neighbors that are mostly from the opposite semantic class (e.g., nonliving). In the present study we investigated the contribution of phonology to orthographic neighborhood effects by comparing effects of phonologically congruent orthographic neighbors (book-hook) to phonologically incongruent orthographic neighbors (sand-wand). The prior presentation of a semantically congruent word produced larger effects on subsequent animacy decisions when the previously presented word was a phonologically congruent neighbor than when it was a phonologically incongruent neighbor. In a second experiment, performance differences between target words with versus without semantically congruent orthographic neighbors were larger if the orthographic neighbors were also phonologically congruent. These results support models of visual word recognition that assume an important role for phonology in cascaded access to meaning.

scholarly journals Schaalvergroting in het syntactische alternantieonderzoek : Een nieuwe analyse van het presentatieve er met automatisch gegenereerde predictoren

2020 ◽  
Vol 25 (1) ◽  
pp. 101-123
Author(s):  
Dirk Speelman ◽  
Stefan Grondelaers ◽  
Benedikt Szmrecsanyi ◽  
Kris Heylen
Keyword(s):  
The Other ◽  
Manual Annotation ◽  
Distributional Analysis ◽  
Syntactic Variation ◽  
Semantic Class ◽  
Semantic Classes ◽  
Reference Corpus ◽  
Lexical Collocations ◽  
The Subject ◽  
Pragmatic Factor

Abstract In this paper, we revisit earlier analyses of the distribution of er ‘there’ in adjunct-initial sentences to demonstrate the merits of computational upscaling in syntactic variation research. Contrary to previous studies, in which major semantic and pragmatic predictors (viz. adjunct type, adjunct concreteness, and verb specificity) had to be coded manually, the present study operationalizes these predictors on the basis of distributional analysis: instead of hand-coding for specific semantic classes, we determine the semantic class of the adjunct, verb, and subject automatically by clustering the lexemes in those slots on the basis of their ‘semantic passport’ (as established on the basis of their distributional behaviour in a reference corpus). These clusters are subsequently interpreted as proxies for semantic classes. In addition, the pragmatic factor ‘subject predictability’ is operationalized automatically on the basis of collocational attraction measures, as well as distributional similarity between the other slots and the subject. We demonstrate that the distribution of er can be modelled equally successfully with the automated approach as in manual annotation-based studies. Crucially, the new method replicates our earlier findings that the Netherlandic data are easier to model than the Belgian data, and that lexical collocations play a bigger role in the Netherlandic than in the Belgian data. On a methodological level, the proposed automatization opens up a window of opportunities. Most important is its scalability: it allows for a larger gamut of alternations that can be investigated in one study, and for much larger datasets to represent each alternation.

scholarly journals Word vs. Class-Based Word Sense Disambiguation

2015 ◽  
Vol 54 ◽  
pp. 83-122 ◽  
Author(s):  
Ruben Izquierdo ◽  
Armando Suarez ◽  
German Rigau
Keyword(s):  
Word Sense Disambiguation ◽  
Coarse Grained ◽  
Semantic Features ◽  
Word Sense ◽  
Simple Method ◽  
Word Meanings ◽  
Semantic Class ◽  
Semantic Classes ◽  
Sense Disambiguation ◽  
Word Senses

As empirically demonstrated by the Word Sense Disambiguation (WSD) tasks of the last SensEval/SemEval exercises, assigning the appropriate meaning to words in context has resisted all attempts to be successfully addressed. Many authors argue that one possible reason could be the use of inappropriate sets of word meanings. In particular, WordNet has been used as a de-facto standard repository of word meanings in most of these tasks. Thus, instead of using the word senses defined in WordNet, some approaches have derived semantic classes representing groups of word senses. However, the meanings represented by WordNet have been only used for WSD at a very fine-grained sense level or at a very coarse-grained semantic class level (also called SuperSenses). We suspect that an appropriate level of abstraction could be on between both levels. The contributions of this paper are manifold. First, we propose a simple method to automatically derive semantic classes at intermediate levels of abstraction covering all nominal and verbal WordNet meanings. Second, we empirically demonstrate that our automatically derived semantic classes outperform classical approaches based on word senses and more coarse-grained sense groupings. Third, we also demonstrate that our supervised WSD system benefits from using these new semantic classes as additional semantic features while reducing the amount of training examples. Finally, we also demonstrate the robustness of our supervised semantic class-based WSD system when tested on out of domain corpus.

The dative alternation in South Asian English(es)

2014 ◽  
Vol 35 (1) ◽  
pp. 7-31 ◽  
Author(s):  
Tobias Bernaisch ◽  
Stefan Th. Gries ◽  
Joybrato Mukherjee
Keyword(s):  
South Asian ◽  
Conditional Inference ◽  
The South ◽  
Double Object ◽  
Indian English ◽  
British English ◽  
Semantic Class ◽  
Dative Alternation ◽  
Syntactic Pattern ◽  
Conditional Inference Trees

The present paper focuses on the modelling of cross-varietal differences and similarities in South Asian English(es) and British English at the level of verb complementation. Specifically, we analyse the dative alternation with GIVE, i.e. the alternation between the double-object construction (John gave Mary a book) and the prepositional dative (John gave a book to Mary) as well as their passivised constructions with regard to the factors that potentially exert an influence on this alternation in seven varieties of English. The South Asian varieties under scrutiny are Bangladeshi English, Indian English, Maldivian English, Nepali English, Pakistani English and Sri Lankan English, while British English serves as the reference variety. The patterns of GIVE are annotated according to the following parameters including potential predictors of the dative alternation: syntactic pattern and semantic class of GIVE; syntactic complexity, animacy, discourse accessibility and pronominality of constituents (cf. Gries 2003b; Bresnan and Hay 2008). The choices of complementation patterns are then statistically modelled using conditional inference trees and a random-forest analysis. The results indicate that many of the predictors found to be relevant in British English are at play in the South Asian varieties, too. The syntactic pattern of GIVE is, in descending order, uniformly influenced by the predictors pronominality of recipient, length of recipient, semantic class of GIVE and length of patient. Interestingly, the predictor country is marginal in accounting for the dative alternation of GIVE across the varieties at hand. Based on this observation, we derive variety-independent protostructions, i.e. abstract combinations of (cross-varietally stable) features with high predictive power for a particular syntactic pattern, which we argue to be part of the lexicogrammatical “common core” (Quirk et al. 1985: 16) of English. The implications of the present paper are twofold. While the order of the predictors regarding their influence on the dative alternation is clearly compatible with earlier studies (cf. e.g. Green 1974; Ransom 1979; Hawkins 1994; Gries 2003b), the stability of the order across varieties of English calls for a) a more fine-grained gradation of linguistic forms and structures at the lexis-grammar interface as indicators of structural nativisation and b) a revision of earlier verb-complementational findings specific to individual or groups of varieties of South Asian English.

scholarly journals Weakly Supervised Conditional Random Fields Model for Semantic Segmentation with Image Patches

2020 ◽  
Vol 10 (5) ◽  
pp. 1679
Author(s):  
Xinying Xu ◽  
Yujing Xue ◽  
Xiaoxia Han ◽  
Zhe Zhang ◽  
Jun Xie ◽  
...  
Keyword(s):  
Random Fields ◽  
Conditional Random Fields ◽  
Semantic Category ◽  
Semantic Segmentation ◽  
Potential Energy Function ◽  
Semantic Class ◽  
Image Patches ◽  
Benchmark Datasets ◽  
Weakly Supervised ◽  
Class Labels

Image semantic segmentation (ISS) is used to segment an image into regions with differently labeled semantic category. Most of the existing ISS methods are based on fully supervised learning, which requires pixel-level labeling for training the model. As a result, it is often very time-consuming and labor-intensive, yet still subject to manual errors and subjective inconsistency. To tackle such difficulties, a weakly supervised ISS approach is proposed, in which the challenging problem of label inference from image-level to pixel-level will be particularly addressed, using image patches and conditional random fields (CRF). An improved simple linear iterative cluster (SLIC) algorithm is employed to extract superpixels. for image segmentation. Specifically, it generates various numbers of superpixels according to different images, which can be used to guide the process of image patch extraction based on the image-level labeled information. Based on the extracted image patches, the CRF model is constructed for inferring semantic class labels, which uses the potential energy function to map from the image-level to pixel-level image labels. Finally, patch based CRF (PBCRF) model is used to accomplish the weakly supervised ISS. Experiments conducted on two publicly available benchmark datasets, MSRC and PASCAL VOC 2012, have demonstrated that our proposed algorithm can yield very promising results compared to quite a few state-of-the-art ISS methods, including some deep learning-based models.

A Semantic Class in the Great Basin

Names ◽  
1978 ◽  
Vol 26 (1) ◽  
pp. 48-57 ◽  
Author(s):  
Thomas L. Clark
Keyword(s):  
Great Basin ◽  
Semantic Class

De la possibilité de suppression du régime des expressions prépositives de concession, comparaison, addition et destination

Revue Romane ◽  
2008 ◽  
Vol 43 (1) ◽  
pp. 151-170
Author(s):  
Silvia Adler
Keyword(s):  
Lexical Status ◽  
Semantic Class

This article pursues a series of studies on French expressions commonly called “prepositional locutions” whose dual goal is, first, the treatment of their lexical status and, second, the possibilities of anaphorisation by a simple suppression of their complement, operation often attributed to the procedure of ellipsis. These studies have concentrated each on the analysis of a specific semantic class of prepositional locutions: spatial, temporal and notional (more precisely, final, causal and consecutive) expressions (cf. Adler, 2006, to appear a, b). The present paper constitutes an additional stage in the puzzle already rather advanced of the analysis of the prepositional locutions, and covers other semantic subclasses: concession, opposition, addition and destination.

Sign in / Sign up

Export Citation Format

Share Document