Single Sentence Research Articles

Machine-to-Machine Visual Dialoguing with ChatGPT for Enriched Textual Image Description

Image captioning is a technique that enables the automatic extraction of natural language descriptions about the contents of an image. On the one hand, information in the form of natural language can enhance accessibility by reducing the expertise required to process, analyze, and exploit remote sensing images, while on the other, it provides a direct and general form of communication. However, image captioning is usually restricted to a single sentence, which barely describes the rich semantic information that typically characterizes remote sensing (RS) images. In this paper, we aim to move one step forward by proposing a captioning system that, mimicking human behavior, adopts dialogue as a tool to explore and dig for information, leading to more detailed and comprehensive descriptions of RS scenes. The system relies on a questions–answers scheme fed by a query image and summarizes the dialogue content with ChatGPT. Experiments carried out on two benchmark remote sensing datasets confirm the potential of such an approach in the context of semantic information mining. Strengths and weaknesses are highlighted and discussed, as well as some possible future developments.

The Impact of Stimuli Length and Analytic Method on Auditory 'Gaydar' Research.

This research investigates voice-based categorization of speakers' sexual orientation, the so-called 'auditory gaydar', while considering stimuli length and the type of measures assessing gaydar judgments. In Study 1 (N=80), heterosexual individuals listened to gay and heterosexual male speakers in short (single word), medium (single sentence), or long (two sentences) recordings. Next, they guess the speakers' sexual orientation on a Kinsey-like and binary choice. Participants were overall inaccurate in recognizing gay speakers when gaydar judgments were provided on a binary choice. Gay speakers were rated as more gay on a Kinsey-like scale than their heterosexual counterparts, but only when short and medium recordings were listened to. Study 2 (N=149) examined gaydar accuracy for both male and female speakers by using the same procedure. Gaydar judgments were overall inaccurate for gay and lesbian speakers. For male speakers, a difference between the perceived sexual orientation of gay and heterosexual speakers emerged when stimuli were long. For female speakers, such a difference occurred only in the short and medium recording conditions. Study 3 (N=137) examined gaydar judgments for male and female speakers when stimuli length was manipulated as the number of words progressively presented in a sentence: short (article), medium (article+first word), long (sentence) stimulus. Overall, gaydar judgments were inaccurate. Gay and lesbian (vs. heterosexual) speakers tended to be rated as more gay on the Kinsey-like scale in the medium stimulus condition. These findings suggest that gaydar judgments can be influenced by the type of measure and stimuli.

Verification Principle and Testability Principle

The paper deals with the conception of logical empiricism developed by Eino Kaila. Eino Kaila, being a thinker close to the Vienna Circle, departs from some of the central ideas of logical positivism. He identifies a limited number of problems in metaphysics that are meaningful and need to be solved, but he declares the rest of metaphysics to be a logical fallacy. For Eino Kaila, it is not the principle of verification (as a criterion of meaning) but the principle of testability that plays the most important role. In addition, he revises the principle of translatability, insisting that it is impossible to translate a single sentence into the language of experience, but it is possible to translate the whole theory to which the sentence belongs. This is related to his structuralist position in the philosophy of science and his understanding of scientific theories as ‘rationalisations’ as opposed to simple inductive generalisations. The paper compares Eino Kaila’s views expressed during the period of his active interaction with the Vienna Circle be regarded as a predecessor of later critics of logical positivism (in particular, W.V.O. Quine).

가능세계의 양상 논리 시스템을 활용한 전설 서사 분석 모델 연구

This study develops a new analysis methodology for legendary narratives based on the possible world theory. As part of this, it develops a narrative analysis model. Unlike “traditional narratology,” possible world theory enables new and diverse approaches related to narrative by establishing various worlds created through stories, and it allows the establishment of related hypotheses and theories. Particularly, it provides a methodology for a new understanding of narratives that are difficult to establish in literature or history, such as legendary narratives.
 Legends are often regarded as fictional narratives that have a certain degree of historicity with temporal and spatial limitations and evidence but are differentiated from history. However, according to White, history is also a fictional narrative that creates a specific world by constructing a plot and indicates a specific ideology or meaning. Finally, historical narratives are simply narratives related to the history of actual events, whereas legendary narratives are non-historical narratives about a world believed to have existed. By breaking away from the conventional perspective of mimesis and assuming various possible worlds, we can plan a new methodology for narrative research by examining the worlds presented by narratives.
 In this discussion, a model for analyzing legend narratives was presented using Doležel's modal logic system. The alethic, deontic, axiological, and epistemic plots were established as segmented units and analysis items that operate in the story world. The modes of each stage of the narrative were analyzed and connected to create meaning in the form of a single sentence. This analysis model was applied to Song Si-yeol legends handed down in the Goesan region to analyze the composition principles of the narrative text and the meaning of the narrative. By applying the secondary narrative analysis model using a modal logic system, it is possible to examine the semantics of the story world within a systematic framework, rather than simply finding the primary indicative meaning or subjective and abstract commentary of the legendary narratives.

Sentence Fusion using Deep Learning

The human process of document summarization involves summarizing a document by sentence fusion. Sentence fusion combines two or more sentences to create an abstract sentence. Sentence fusion is useful to convert an extractive summary to an abstractive summary. The extractive summary contains a set of salient sentences selected from a single document or multiple related documents. Redundancy creates problems while creating an extractive summary because it contains sentences whose segments or phrases are redundant. Sentence fusion helps to remove redundancy by fusing sentences into a single abstract sentence. This moves an extractive summary to an abstractive summary. In this paper, we present an approach that uses a deep learning model for sentence fusion. which is trained over a large dataset. We have tested our approach through both manual evaluation and system evaluation. The result of our proposed approach shows that our model is good enough to fuse sentences effectively.

Hong Kong Corpus of Chinese Sentence and Passage Reading

Recent years have witnessed a mushrooming of reading corpora that have been built by means of eye tracking. This article showcases the Hong Kong Corpus of Chinese Sentence and Passage Reading (HKC for brevity), featured by a natural reading of logographic scripts and unspaced words. It releases 28 eye-movement measures of 98 native speakers reading simplified Chinese in two scenarios: 300 one-line single sentences and 7 multiline passages of 5,250 and 4,967 word tokens, respectively. To verify its validity and reusability, we carried out (generalised) linear mixed-effects modelling on the capacity of visual complexity, word frequency, and reading scenario to predict eye-movement measures. The outcomes manifest significant impacts of these typical (sub)lexical factors on eye movements, replicating previous findings and giving novel ones. The HKC provides a valuable resource for exploring eye movement control; the study contrasts the different scenarios of single-sentence and passage reading in hopes of shedding new light on both the universal nature of reading and the unique characteristics of Chinese reading.

Syntactic Expressivity in Cultural Reporting: Insights from Australian Media

The article aims to study the usage and influence of expressive syntax within a variety of culture-related articles featured in contemporary Australian digital media. Leveraging textual analysis on online articles sourced from prominent Australian newspapers such as “The Sydney Morning Herald”, “The Alice Springs News”, “Katherine Times”, and “The New Daily”, the research provides insights into the frequency and novelty of expressive syntactic elements employed within. Particularly prevalent techniques, such as parenthetical inclusions, interrogative sentences, elliptical, nominative, and parcelled constructions, are thoroughly examined for their expressive potential. The study offers a comprehensive breakdown of these syntactical elements, highlighting key types, delineating their general attributes, and enumerating the specifics of their implementation within the researched texts. It especially highlights the frequent interaction of various expressive syntax within single sentences, proposing that this phenomenon increases the techniques’ effectiveness. Furthermore, the study explores the amplification of expression through the usage of specific punctuation marks, including quotation marks, dashes, colons, brackets, semicolons, and ellipses, each undergoing a thorough analysis for potential effect. In summarizing the principal outcomes of the research, the study asserts the significant role of expressive syntactic techniques in crafting engaging and impactful cultural articles within modern Australian media.

The Journey from a Single Sentence to a Well-Written Essay

The Journey from a Single Sentence to a Well-Written Essay

Exploiting global contextual information for document-level named entity recognition

Named entity recognition (NER, also known as entity chunking/extraction) is a fundamental sub-task of information extraction, which aims at identifying named entities from an unstructured text into pre-defined classes. Most of the existing works mainly focus on modeling local-context dependencies in a single sentence for entity type prediction. However, they may neglect the clues derived from other sentences within a document, and thus suffer from the sentence-level inherent ambiguity issue, which may make their performance drop to some extent. To this end, we propose a Global Context enhanced Document-level NER (GCDoc) model for NER to fully exploit the global contextual information of a document in different levels, i.e., word-level and sentence-level. Specifically, GCDoc constructs a document graph to capture the global dependencies of words for enriching the representations of each word in word-level. Then, it encodes the adjacent sentences for exploring the contexts across sentences to enhance the representation of the current sentence via the specially devised attention mechanism. Extensive experiments on two benchmark NER datasets (i.e., CoNLL 2003 and Onenotes 5.0 English dataset) demonstrate the effectiveness of our proposed model, as compared to the competitive baselines.

On the Concept of Translation Unit in a Machine Translation Framework

The article looks at the question whether the concept of translation unit might apply to the sphere of machine translation and whether the size of the unit influences the quality of translation. While modern machine translation systems offer an acceptable level of quality, a number of problems mainly related to the structural organization of the text remain unresolved, hence the question posed in the paper. The article offers a review of modern readings of the concept and pays special attention to the question whether the scope of the term changes depending on whether the object of research is the target text or the translation process. The paper also provides a quick look on the research methods for both text-oriented and process-oriented approaches, such as comparative analysis of language pairs and Think Aloud Protocol. Based on a review of existing machine translation models, each of them is analyzed to answer the question whether a unit of translation can be defined for a given system and what its size is. It is concluded that a unit of translation can be viewed as either a unit of analysis or a unit of processing with respect to text-oriented and process-oriented perspectives on to the study of translation. The unit of translation has a dynamic character and influences the quality of the target text. In machine translation, the unit of translation as a unit of analysis is not applicable for systems based on probabilistic non-linguistic methods. For rule-based machine translation systems, both readings of the unit of translation concept are applicable, but hardly go beyond a single sentence. Accordingly, at least one type of translation problem – intra-textual relations resolutions – remains largely unaddressed in the present state of affairs in machine translation.

A NEW ACROSTIC AND TELESTIC AT LAVS PISONIS 227–30?

AbstractThis article proposes a new acrostic (SAPI) and telestic (SOIS) at Laus Pisonis 227–30. Their position opposite one another is an indication that they are to be read as a single sentence and an admonition to both dedicatee and reader that poet and patron need each other to gain eternal fame. The telestic allows us to reconstruct the poet's usus scribendi of the reflexive possessive pronoun suus.

White Matter Correlates of Speech in Cerebrovascular Disease

AbstractBackgroundAnalyses of spoken discourse (speech beyond single words or sentences) can detect cognitive impairment in cerebrovascular disease (CVD) [Roberts, et al. (2021). Top Lang Disord 41(1):73‐98]. However, the neurological basis for altered spoken discourse in CVD is poorly defined. This study examined the association between spoken discourse and visible white matter damage detected using structural magnetic resonance imaging (MRI) and potential early‐stage damage in normal‐appearing white matter (NAWM) detected using diffusion tensor imaging (DTI) to better define the link between CVD‐related neurodegeneration and altered speech.MethodSpoken discourse and 3T MRI data were obtained from the CVD cohort of the Ontario Neurodegenerative Disease Research Initiative (n = 127). Spoken discourse analyses were completed previously [Roberts et al., 2021]. White matter hyperintensities (WMH) and the NAWM portions of the inferior longitudinal fasciculus (ILF), superior longitudinal fasciculus – parietal (SLFp) and temporal (SLFt) endings, and uncinate fasciculus (UNC) were segmented in each hemisphere (Figure 1). Mean values of fractional anisotropy (FA) and mean diffusivity (MD) within NAWM, and normalized WMH volume were calculated for each of the white matter tracts. Three canonical correlation analyses examined associations between 10 spoken discourse measures and the imaging metrics (WMH volume, FA, MD) in each tract.ResultCanonical correlations were significant for FA (rc = 0.51, p<0.05) and MD (rc = 0.56, p<0.05) but not WMH volume (rc = 0.42, p = 0.69). Among the white matter tracts, the strongest canonical loadings were seen for FA and MD in the left SLFp (FA: 0.80; MD: ‐0.61) and left SLFt (FA: 0.60; MD: ‐0.40) (Figure 2). Higher FA and lower MD in the left SLFp and left SLFt were associated with fluency, information content, and syntax performance.ConclusionSpoken discourse performance was associated with microstructural integrity of NAWM but not visible white matter damage. These results suggest that DTI metrics provide a more relevant assessment of white matter tract function in relation to speech than a volumetric measure of tract lesion load. Of the white matter tracts investigated in this study, spoken discourse performance in CVD was most strongly linked to white matter microstructure in the parietal and temporal endings of the left superior longitudinal fasciculus.

White Matter Correlates of Speech in Cerebrovascular Disease

AbstractBackgroundAnalyses of spoken discourse (speech beyond single words or sentences) can detect cognitive impairment in cerebrovascular disease (CVD) [Roberts, et al. (2021). Top Lang Disord 41(1):73‐98]. However, the neurological basis for altered spoken discourse in CVD is poorly defined. This study examined the association between spoken discourse and visible white matter damage detected using structural magnetic resonance imaging (MRI) and potential early‐stage damage in normal‐appearing white matter (NAWM) detected using diffusion tensor imaging (DTI) to better define the link between CVD‐related neurodegeneration and altered speech.MethodSpoken discourse and 3T MRI data were obtained from the CVD cohort of the Ontario Neurodegenerative Disease Research Initiative (n = 127). Spoken discourse analyses were completed previously [Roberts et al., 2021]. White matter hyperintensities (WMH) and the NAWM portions of the inferior longitudinal fasciculus (ILF), superior longitudinal fasciculus – parietal (SLFp) and temporal (SLFt) endings, and uncinate fasciculus (UNC) were segmented in each hemisphere (Figure 1). Mean values of fractional anisotropy (FA) and mean diffusivity (MD) within NAWM, and normalized WMH volume were calculated for each of the white matter tracts. Three canonical correlation analyses examined associations between 10 spoken discourse measures and the imaging metrics (WMH volume, FA, MD) in each tract.ResultCanonical correlations were significant for FA (rc = 0.51, p<0.05) and MD (rc = 0.56, p<0.05) but not WMH volume (rc = 0.42, p = 0.69). Among the white matter tracts, the strongest canonical loadings were seen for FA and MD in the left SLFp (FA: 0.80; MD: ‐0.61) and left SLFt (FA: 0.60; MD: ‐0.40) (Figure 2). Higher FA and lower MD in the left SLFp and left SLFt were associated with fluency, information content, and syntax performance.ConclusionSpoken discourse performance was associated with microstructural integrity of NAWM but not visible white matter damage. These results suggest that DTI metrics provide a more relevant assessment of white matter tract function in relation to speech than a volumetric measure of tract lesion load. Of the white matter tracts investigated in this study, spoken discourse performance in CVD was most strongly linked to white matter microstructure in the parietal and temporal endings of the left superior longitudinal fasciculus.

Designing remote synchronous auditory comprehension assessment for severely impaired individuals with aphasia.

The use of telepractice in aphasia research and therapy is increasing in frequency. Teleassessment in aphasia has been demonstrated to be reliable. However, neuropsychological and clinical language comprehension assessments are not always readily translatable to an online environment and people with severe language comprehension or cognitive impairments have sometimes been considered to be unsuitable for teleassessment. This project aimed to produce a battery of language comprehension teleassessments at the single word, sentence and discourse level suitable for individuals with moderate-severe language comprehension impairments. Assessment development prioritised response consistency and clinical flexibility during testing. Teleassessments were delivered in PowerPoint over Zoom using screen sharing and remote control functions. The assessments were evaluated in 14 people with aphasia and 9 neurotypical control participants. Modifiable assessment templates are available here: https://osf.io/r6wfm/. People with aphasia were able to engage in language comprehension teleassessment with limited carer support. Only one assessment could not be completed for technical reasons. Statistical analysis revealed above chance performance in 141/151 completed assessments. People with aphasia, including people with moderate-severe comprehension impairments, are able to engage with teleassessment. Successful teleassessment can be supported by retaining clinical flexibility and maintaining consistent task demands. What is already known on the subject Teleassessment for aphasia is reliable but assessment of auditory comprehension is difficult to adapt to the online environment. There has been limited evaluation of the ability of people with severe aphasia to engage in auditory comprehension teleassessment. What this paper adds to existing knowledge Auditory comprehension assessment can be adapted for videoconferencing administration while maintaining clinical flexibility to support people with severe aphasia. What are the potential or actual clinical implications of this work? Teleassessment is time and cost effective and can be designed to support inclusion of severely impaired individuals.

Sentence Semantic Similarity based Complex Network approach for Word Sense Disambiguation

Word Sense Disambiguation is a branch of Natural Language Processing(NLP) that deals with multi-sense words. The multi-sense words are referred to as the polysemous words. The term lexical ambiguity is introduced by the multi-sense words. The existing sense disambiguation module works effectively for single sentences with available context information. The word embedding plays a vital role in the process of disambiguation. The context-dependent word embedding model is used for disambiguation. The main goal of this research paper is to disambiguate the polysemous words by considering available context information. The main identified challenge of disambiguation is the ambiguous word without context information. The discussed complex network approach is disambiguating ambiguous sentences by considering the semantic similarities. The sentence semantic similarity-based network is constructed for disambiguating ambiguous sentences. The proposed methodology is trained with SemCor, Adaptive-Lex, and OMSTI standard lexical resources. The findings state that the discussed methodology is working fine for disambiguating large documents where the sense of ambiguous sentences is on the adjacent sentences.

Event Detection with Word-Word Relation Classification

Event detection (ED) is a critical task in information extraction, aiming to identify triggers and types. Current research has increasingly focused on fine-grained types, where a single sentence may contain multiple trigger words and event types. Previous models only considered sentence-level features, neglecting word-word features and word positional information in the text. We propose a novel labeling scheme that treats event detection as a word-word relation recognition task. In this approach, we first identify relationships between word pairs and then utilize these relationships to perform trigger and event-type detection. By adopting this method, we can efficiently and concurrently identify triggers and types in a sentence. Leveraging word pair relationships effectively addresses scenarios in which multiple trigger words appear in a single sentence. The results demonstrate that our approach outperforms several baseline models.

On the instability of further pre-training: Does a single sentence matter to BERT?

We observe a remarkable instability in BERT-like models: minimal changes in the internal representations of BERT, as induced by one-step further pre-training with even a single sentence, can noticeably change the behaviour of subsequently fine-tuned models. While the pre-trained models seem to be essentially the same, also by means of established similarity assessment techniques, the measurable tiny changes appear to substantially impact the models’ tuning path, leading to significantly different fine-tuned systems and affecting downstream performance. After testing a very large number of combinations, which we briefly summarize, the experiments reported in this short paper focus on an intermediate phase consisting of a single-step and single-sentence masked language modeling stage and its impact on a sentiment analysis task. We discuss a series of unexpected findings which leave some open questions over the nature and stability of further pre-training.

A Corpus-Based Study of the Order of Adjectives in Japanese Antonym Sequences: The Role of Markedness and Frequency

Abstract This study investigates whether the order of adjectives in antonym pairs (e.g., ii ‘good, warui ‘bad’) in single sentences of Japanese follows principles of markedness and whether the more frequently used antonym is more likely to be used first. Based on data collected from the Tsukuba Web Corpus, the results of the study were that both markedness and frequency equally influence the order of antonym pairs. This study suggests that markedness plays a role in determining the ordering of adjectives in antonym pairs. Additionally, it was found that markedness affects the order in terms of factors of positivity and magnitude.

Biology in a social context: a comprehensive analysis of humanization in introductory biology textbooks

To grapple with the sterility and Whiteness of Western science, scholars have proposed a pedagogical shift to culturally relevant and/or culturally sustaining pedagogy. A key tenet of culturally relevant pedagogy is a focus on developing students’ ability to use the knowledge they obtain to identify, analyze, and solve real-world problems. Thus, the ability to foster this consciousness among students and make justice/injustice visible within biology curricula is an act of humanization. Here, we characterize and quantify the extent to which six prominent introductory biology US-based textbooks include humanizing content. First, we built consensus on what it means to humanize biology in a textbook by iteratively revising a coding protocol until we achieved a continuum of humanization. Our continuum evaluates the quantity, location, and the nature of the humanizing element within the textbook. Then, we used the continuum to collect data through qualitative coding: each chapter of each textbook was coded by two coders who came to consensus on the humanizing elements within. We find that in general, the inclusion of humanizing content in introductory biology textbooks is rare: of the 9,670 pages of textbooks that we analyzed, we found 1,352 humanizing passages but the vast majority of these were discussed in a single sentence (23%) or multiple sentences (61%), rarely multiple paragraphs (13%) or entire sections (2%). Similarly, of the 9,262 questions in the books (e.g., in section or chapter summaries), only 2.5% of them were humanizing and of those, only (64%) provided an answer, and of the ones that provided an answer, we only coded 42% of the answers as humanizing. In addition to quantifying the amount of humanization, we also describe the ways in which the passages were presented. For example, only about 9% of the humanizing passages included nuance, 5% discussed equity/inequity, and only 4% positioned biology as a means to accomplish justice. In all, we present what we believe is the most comprehensive assessment of humanizing elements in introductory biology textbooks and pair that with specific guidance to instructors who seek to include humanizing elements in their classes.

Teaching Medical Students to Use Supercomputers: A Personal Reflection.

At the "Kick Off" meeting for CompBioMed (compbiomed.eu), which was first funded in October 2016, I had no idea that one single sentence ("I wish I could teach this to medical students") would lead to a dedicated program of work to engage the clinicians and biomedical researchers of the future with supercomputing. This program of work which, within the CompBiomed Centre of Excellence, we have been calling "the CompBioMed Education and Training Programme," is a holistic endeavor that has been developed by and continues to be delivered with the expertise and support from experimental researchers, computer scientists, clinicians, HPC centers, and industrial partners within or associated with CompBioMed. The original description of the initial educational approach to training has previously been published (Townsend-Nicholson Interface Focus 10:20200003, 2020). In this chapter, I describe the refinements to the program and its delivery, emphasizing the highs and lows of delivering this program over the past 6 years. I conclude with suggestions for feasible measures that I believe will help overcome the barriers and challenges we have encountered in bringing a community of users with little familiarity of computing beyond the desktop to the petascale and beyond.