Semantic Variance Research Articles

CPR++: Object Localization via Single Coarse Point Supervision.

Point-based object localization (POL), which pursues high-performance object sensing under low-cost data annotation, has attracted increased attention. However, the point annotation mode inevitably introduces semantic variance due to the inconsistency of annotated points. Existing POL heavily rely on strict annotation rules, which are difficult to define and apply, to handle the problem. In this study, we propose coarse point refinement (CPR), which to our best knowledge is the first attempt to alleviate semantic variance from an algorithmic perspective. CPR reduces the semantic variance by selecting a semantic centre point in a neighbourhood region to replace the initial annotated point. Furthermore, We design a sampling region estimation module to dynamically compute a sampling region for each object and use a cascaded structure to achieve end-to-end optimization. We further integrate a variance regularization into the structure to concentrate the predicted scores, yielding CPR++. We observe that CPR++ can obtain scale information and further reduce the semantic variance in a global region, thus guaranteeing high-performance object localization. Extensive experiments on four challenging datasets validate the effectiveness of both CPR and CPR++. We hope our work can inspire more research on designing algorithms rather than annotation rules to address the semantic variance problem in POL.

Emotion Analysis and Dialogue Breakdown Detection in Dialogue of Chat Systems Based on Deep Neural Networks

In dialogues between robots or computers and humans, dialogue breakdown analysis is an important tool for achieving better chat dialogues. Conventional dialogue breakdown detection methods focus on semantic variance. Although these methods can detect dialogue breakdowns based on semantic gaps, they cannot always detect emotional breakdowns in dialogues. In chat dialogue systems, emotions are sometimes included in the utterances of the system when responding to the speaker. In this study, we detect emotions from utterances, analyze emotional changes, and use them as the dialogue breakdown feature. The proposed method estimates emotions by utterance unit and generates features by calculating the similarity of the emotions of the utterance and the emotions that have appeared in prior utterances. We employ deep neural networks using sentence distributed representation vectors as the feature. In an evaluation of experimental results, the proposed method achieved a higher dialogue breakdown detection rate when compared to the method using a sentence distributed representation vectors.

What it’s like to be a _____: why it’s (often) unethical to use VR as an empathy nudging tool

In this article, we apply the literature on the ethics of choice-architecture (nudges) to the realm of virtual reality (VR) to point out ethical problems with using VR for empathy-based nudging. Specifically, we argue that VR simulations aiming to enhance empathic understanding of others via perspective-taking will almost always be unethical to develop or deploy. We argue that VR-based empathy enhancement not only faces traditional ethical concerns about nudges (autonomy, welfare, transparency), but also a variant of the semantic variance problem that arises for intersectional perspective-taking. VR empathy simulations deceive and manipulate their users about their experiences. Despite their often laudable goals, such simulations confront significant ethical challenges. In light of these goals and challenges, we propose VR designers shift from designing simulations aimed at producing empathic perspective-taking to designing simulations aimed at generating sympathy for their targets. These simulations, we claim, can avoid the most serious ethical issues associated with VR nudges, semantic variance, and intersectionality.

Identidade e variação semânticas do verbo SECAR no português do Brasil

Resultado da pesquisa realizada por Kiihl (2016) e Romero (2016), objetiva-se mostrar neste artigo como a identidade semântica verbal se elabora na integração do verbo ao enunciado. A Teoria das Operações Predicativas e Enunciativas, referencial teórico que fundamenta o quadro analítico-metodológico do trabalho, concebe o sentido verbal como o produto de seu enunciar, o que implica ser a identidade semântica da unidade, quando fora da contextualização que lhe confere o enunciado, da ordem de um esquema operatório invariante, denominado forma esquemática. O estudo, direcionado ao verbo SECAR em PB, delimita o referido esquema por meio da análise de enunciados variados, extraídos de fontes primárias e secundárias, mostrando como ele é o responsável por organizar a variação semântica verbal.

Some examples from a corpus investigation of potentially confusing polysemy and false friends in the terminology of music, centred on English and French, with equivalents in five other languages

This paper presents five of the initial items of a semantic study of confusable terms and false friends present in the lexis of music in English and French. All forms are also given in Italian, German, Spanish, Russian and Chinese. The object of study is a series of texts from corpora provided by New Sketch Engine, a product of Lexical Computing Limited. In the commentary section, the semantic variance between the different items within the same topic is studied. The approach is essentially synchronic, with priority being given to the examination of current usage. Diachronic considerations are also included where it has been possible to explain, or at least hypothesize how the meaning of the various terms has evolved. The main practical application of the paper is to provide a tool for reducing ambiguity in the area of music terminology. It will be of potential interest to all those concerned by the study and practice of music and the language of music in countries where the languages referred to in this study are in use.

Semantic variance: An intuitive measure for ontology accuracy evaluation

Ontology evaluation is a relevant issue in the field of knowledge representation. It aims at quantifying the quality of ontologies, so that potential users can have an idea of their accuracy and thereby select the most appropriate ontology for a specific application. Many of the ontology evaluation methods and frameworks available in the literature assess the quality of ontologies according to their structural features, even though most of these methods propose ad hoc aggregations of such features that lack a theoretical basis. Inspired by recent empirical studies showing that some structural features are better suited to predict the semantic accuracy of ontologies, we present in this paper the notion of semantic variance of an ontology. Semantic variance is an intuitive and inherently semantic measure to evaluate the accuracy of ontologies. Unlike ad hoc methods, our proposal is a mathematically coherent extension of the standard numerical variance to measure the semantic dispersion of the taxonomic structure of ontologies. In our experiments performed over a set of widely used ontologies, the proposed semantic variance positively correlated with the structural features of ontologies that best predicted their accuracy in previous studies. Moreover, our measure also provided a good prediction of the ontological accuracy in one of the most essential knowledge-based tasks: assessing the semantic similarity between concepts. These results suggest that the semantic variance can be used as a generic, quantitative and theoretically coherent score to evaluate the accuracy of ontologies.

Towards a universal detector by mining concepts with small semantic gaps

Can we have a universal detector that could visually recognize unseen objects with no training exemplars available? Such a detector is so desirable, as there are hundreds of thousands of object concepts in human vocabulary but few labeled image examples available. In this study, we attempt to build such a universal detector to predict concepts in the absence of training data. First, by considering both semantic relatedness and visual variance, we mine a set of realistic small-semantic-gap (SSG) concepts from a large-scale image corpus, i.e., ImageNet, which comprises 4961 concepts and nearly 4 million images. The discovered SSG concepts can be depicted well by visual models and their detectors can deliver reasonably satisfactory recognition accuracies. From these distinctive visual models, we then leverage the semantic ontology knowledge and co-occurrence statistics of concepts to extend visual recognition to unseen concepts. The rational is that object concepts generally co-occur in a real-life image. Their visual co-occurrence and semantic ontology provide the possibility for concept recognition to transcend the visual learning of image exemplars, and therefore, enable the detector to predict unseen realistic concepts without training samples. To the best of our knowledge, this work presents the first research attempting to substantiate the semantic gap measuring of a large amount of concepts and leverage visually learnable concepts to predicate those with no training images available. Testings on NUS-WIDE dataset demonstrate that the selected concepts with small semantic gaps can be well modeled and the prediction of unseen concepts delivers promising results with comparable accuracy to preliminary training-based methods.

Competence and Trust in Choice Architecture

Richard Thaler and Cass Sunstein’s Nudge advances a theory of how designers can improve decision-making in various situations where people have to make choices. We claim that the moral acceptability of nudges hinges in part on whether they can provide an account of the competence required to offer nudges, an account that would serve to warrant our general trust in choice architects. What needs to be considered, on a methodological level, is whether they have clarified the competence required for choice architects to prompt subtly our behaviour toward making choices that are in our best interest from our own perspectives. We argue that, among other features, an account of the competence required to offer nudges would have to clarify why it is reasonable to expect that choice architects can understand the constraints imposed by semantic variance. Semantic variance refers to the diverse perceptions of meaning, tied to differences in identity and context, that influence how users interpret nudges. We conclude by suggesting that choice architects can grasp semantic variance if Thaler and Sunstein’s approach to design is compatible with insights about meaning expressed in science and technology studies and the philosophy of technology.

Mapping semantic information in virtual space: dimensions, variance and individual differences

This paper reports two studies investigating the computer-based representation of the semantic information content of databases using object location in two- and three-dimensional virtual space. In the first study, the cognitive demands associated with performing an information search task were examined under conditions where the “goodness of fit” of the spatial-semantic “mapping” was manipulated. The effects of individual differences in spatial ability and associative memory ability also were considered. Results indicated that performance equivalence, between two- and three-dimensional interfaces, could be achieved when the two-dimensional interface accounted for between 50 and 70% of the semantic variance accounted for by the three-dimensional solution. A second study, in which automatic text analysis was used to generate two- and three-dimensional solutions for document sets of varying sizes and types, supported the conclusion that, for the purpose of information search, the amount of additional semantic information that can be conveyed by a three-dimensional solution does not outweigh the associated additional cognitive demands.

Variability in Test Scores of Forms A and B of the Peabody Picture Vocabulary Test Derived from Sampling a First-Grade Population

Previous studies have suggested that young children’s performances on the Peabody Picture Vocabulary Test vary considerably. These data infer similar performance variability in all populations. To determine the credibility of this semantic variance, Forms A and B of the PPVT were administered to 53 first-grade children in the Pittsburgh, Pennsylvania area. The results provided an additional demonstration of the variability between Forms A and B of the PPVT. Response data from previous studies were further analyzed revealing statistically significant differences between means of the group raw scores.

Semantic Variance in Recommendations of Student Potential

This study investigates clarity of communication among high school counselors and university admissions officers in their use of selected terminology to describe three types of scholars: potentially low‐, average‐, and exceptional‐achieving scholars. A “Descriptive Word Scale” was designed to investigate the degree to which semantic variance was present in the selected terminology used to describe these three types of scholars among counselors, among admissions officers, and between the two groups. A sample of 300 counselors and admissions officers was drawn. Basic theoretical assumptions were drawn from the theories of Miller, Hackett, et al., and Schramm. Major findings included significant variability among counselors and among admissions officers and between the groups for certain terminology. It was possible to distinguish which of the terminology might be more dependable for attaining clarity of communication.