Search results

To understand the meaning of a sentence, humans can focus on important words in the sentence, which reflects our eyes staying on each word in different gaze time or times. Thus, some studies utilize eye-tracking values to optimize the attention mechanism in deep learning models. But these studies lack to explain the rationality of this approach. Whether the attention mechanism possesses this feature of human reading needs to be explored.

The authors conducted experiments on a sentiment classification task. Firstly, they obtained eye-tracking values from two open-source eye-tracking corpora to describe the feature of human reading. Then, the machine attention values of each sentence were learned from a sentiment classification model. Finally, a comparison was conducted to analyze machine attention values and eye-tracking values.

Through experiments, the authors found the attention mechanism can focus on important words, such as adjectives, adverbs and sentiment words, which are valuable for judging the sentiment of sentences on the sentiment classification task. It possesses the feature of human reading, focusing on important words in sentences when reading. Due to the insufficient learning of the attention mechanism, some words are wrongly focused. The eye-tracking values can help the attention mechanism correct this error and improve the model performance.

Our research not only provides a reasonable explanation for the study of using eye-tracking values to optimize the attention mechanism but also provides new inspiration for the interpretability of attention mechanism.

Developing attention-aware systems and interfaces based on eye tracking technology could revolutionize mainstream human–computer interaction to make the interaction between human beings and computers more intuitive, effective and immersive than can be achieved traditionally using a computer mouse. This paper aims to propose an eye-controlled interactive reading system (ECIRS) that uses human eyes instead of the traditional mouse to control digital text to support screen-based digital reading.

This study uses a quasi-experimental design to examine the effects of an experimental group and a control group of learners who, respectively, used the ECIRS and a mouse-controlled interactive reading system (MCIRS) to conduct their reading of two types of English-language text online – pure text and Q&A-type articles on reading comprehension, cognitive load, technology acceptance, and reading behavioural characteristics. Additionally, the effects of learners with field-independent (FI) and field-dependence (FD) cognitive styles who, respectively, used the ECIRS and MCIRS to conduct their reading of two types of English-language text online – pure text and Q&A-type articles on reading comprehension are also examined.

Analytical results reveal that the reading comprehension of learners in the experimental group significantly exceeded those in the control group for the Q&A article, but the difference was insignificant for the pure text article. Moreover, the ECIRS improved the reading comprehension of field-independent learners more than it did that of field-dependent learners. Moreover, neither the cognitive loads of the two groups nor their acceptance of the technology differed significantly, whereas the reading time of the experimental group significantly exceeded that of the control group. Interestingly, for all articles, the control group of learners read mostly from top to bottom without repetition, whereas most of the learners in the experimental group read most paragraphs more than once. Clearly, the proposed ECIRS supports deeper digital reading than does the MCIRS.

This study proposes an emerging ECIRS that can automatically provide supplementary information to a reader and control a reading text based on a reader’s eye movement to replace the widely used mouse-controlled reading system on a computer screen to effectively support digital reading for English language learning. The implications of this study are that the highly interactive reading patterns of digital text with ECIRS support increase motivation and willingness to learn while giving learners a more intuitive and natural reading experience as well as reading an article online with ECIRS support guides learners’ attention in deeper digital reading than does the MCIRS because of simultaneously integrating perceptual and cognitive processes of selection, awareness and control based on human eye movement.

Human libraries are special libraries that give access to real persons known as “human books.” Readers “check out” human books by attending reading sessions organized by the human library. The majority of previous studies focus on reporting the achievement of human libraries, especially on reducing prejudice and promoting social inclusion. There is a dearth of research to investigate how a human library serves and interacts with society. In this regard, this study aims at examining the operation, development, opportunities and challenges of the New Taipei City Human Library (NTCHL), the first and largest human library in Taiwan.

This study adopts a triangulation methodology to collect and analyze a multitude of data. Participatory observation intends to understand the operation of the NTCHL. In-depth interview is conducted to collect the opinions of the staff of the NTCHL to identify the challenges and opportunities of the human library. Content analysis on relevant documents is performed to review the development and performance of the NTCHL.

Results indicate that the NTCHL has successfully developed a good collection of human books. The NTCHL not only provides an alternative channel of learning, but also helps build connections between community members. The number of book loans and readers has substantially increased since NTCHL' inception, indicating an increasing demand for human library services in Taiwan. While the NTCHL strives for providing readers with better and diversified services, the issues of limited involvement and commitment of human books, heavy workload and the difficulty to establish a good collection of human books need to be addressed.

This study investigates a human library in Taiwan, but human libraries in different places are different in some aspects because those human libraries' social contexts are different. Future studies are recommended to compare the similarities and differences between different human libraries systematically. Furthermore, this study identifies the benefits of the human library but human library's effectiveness is not quantified and evaluated. Future studies may conduct quantitative assessments to fill the research gap.

Human libraries must strive for providing better and diversified services to readers, but there are implications for workload and administration. The experience of the NTCHL may cast light on the future development of human libraries in Taiwan and other countries around the world.

This study shows that the NTCHL has brought a couple of benefits to the society of Taiwan. First, the NTCHL provides readers with an alternative experience of reading. This is an effective way to transmit informal or tacit knowledge. Second, building connections between community members was another benefit of the NTCHL. As modern people are becoming more accustomed to virtual life on the internet which suppresses interpersonal communication, the NTCHL represents an effective platform for communication and sharing of interpersonal experiences.

The novelty of this study is the use of triangulation methodology to investigate the operation, development, opportunities and challenges of the first and largest human library in Taiwan. The findings of this study may cast light on the future development of human libraries in Taiwan and other countries.

Several aspects of the e‐book revolution are reviewed, as well as some related issues confronting libraries. Regardless of format, texts and text‐bearing devices have relationships of mutual dependence, and readers simultaneously experience both. The dominant relationship between texts and text‐bearing devices is shifting from static to dynamic. The e‐book revolution is more about new distribution systems for content, new digital rights management systems, and perhaps an unwitting or inchoate power struggle among the principal interested parties, than it is about the design and diffusion of dedicated reading devices. The e‐book revolution opens up possibilities for new and improved post‐retrieval processing of texts, defined as anything a person can do with a text after it has been retrieved. Librarians need to reassert – especially to the fledgling e‐book industry – the enduring principle of libraries as a social good. The two biggest challenges facing libraries are how to make the transition to an era dominated by dynamic relationships between texts and text‐bearing devices, and how to foster and facilitate robust and complex post‐retrieval processing of texts.

Since the change to a democratic society and government in South Africa, lecturers and students in higher education have collaborated with community partners to establish sustainable initiatives to enhance students’ social responsibility and benefit communities. This chapter shares insights on a collaborative service-learning project where different arts-based literacies, including art, reading, poetry, human movement and writing were used to enhance reading and writing, culminating in an annual interactive Community Engagement Day on the campus of a higher education institution. The day was organized by a team of staff and students and attended by 50 learners and three teachers from three schools in the area. The chapter describes different stations and activities which included topics such as safety, human movement programs, reading and writing activities and writing of poems. The art lecturer and her team helped each learner to paint a “feather,” culminating in the theme of the day, which was to “spread your wings.” The learners and students completed evaluation forms after their experiences, followed by interviews with lecturers. Data show the value of the day for all participants, emphasizing collaboration across faculties. It shows that true, integrated effective community engagement is built on reciprocal partnerships and collaborative service-learning projects.

This study aims to investigate “immersive reading,” which occurs when individuals read text while in a virtual reality (VR) or augmented reality (AR) environment.

In Experiment 1, 64 participants read text passages and answered multiple-choice questions in VR and AR head-mounted displays (HMDs) compared with doing the same task on liquid crystal display (LCD). In Experiment 2, 31 participants performed the same reading tasks but with two VR HMDs of different display quality.

Compared with reading on LCD as the baseline, participants reading in VR and AR HMDs got 82% (VR) and 88% (AR) of the information accurately. Participants tended to respond more accurately and faster, though not statistically significant, with the VR HMD of higher pixel density in the speed-reading task.

The authors observed the speed and accuracy of reading in VR and AR environments, compared with the reading speed and accuracy on an LCD monitor. The authors also compared the reading performance on two VR HMDs that differed in display quality but were otherwise similar in every way.

The purpose of this paper is to analyse the accounting research project concerned with accounting narrative obfuscation, focusing on the translation of the concept of readability from educational psychology via an earlier literature concerned with the readability of accounting narratives per se.

This paper uses actor-network theory and examines, in particular, the need for a network to accommodate the interests of its actors and the consequent risk of failure.

The analysis shows that the project is failing because the network seeking to support it is failing, and failing because of its inability to adapt sufficiently to accommodate the interests of its constituents. This failure is contrasted with the earlier concern with readability per se, which did see a successful reconfiguration of actors’ interests.

The puzzle of the maladjustment of the network concerned with obfuscation is examined and it is suggested that it is a consequence of interests prevailing in the wider academic research network within which the relevant human actors are embedded.

The reasons for the failure of the project are bound up in the wider circumstances of the contemporary accounting research community and may affect scholars’ capacity to pursue knowledge effectively.

This paper contributes to a modest stream of actor–network analysis directed at accounting research itself.

This article combines Mead's notion of sociality with his implicit theory of morality. Specifically, it uses Mead's emphasis on temporality to analyze decisions made by key characters in the cinematic adaptation (Amazon TV) of Philip Dick's novel, The Man in the High Castle. Using a selective and subversive method to read into this adaptation, I regard Mead's view of morality as complex and as distinguishing between a morality in the specious present and a morality grounded in sociality. The paper links Mead and Mead's pragmatic emphasis to varieties of characters representing immoral foils (e.g., Nazis) and everyday lives to show how morality can emerge from a variety of standpoints, locating Mead's position as distinct from moral absolutism and moral relativity.