Reading Comprehension Models Research Articles

A Dialogue State Tracking Model with Slot Embedding

Dialogue State Tracking (DST) is the core component of the task-oriented dialogue systems. Although recent neural DST models have made great progress, they often ignore the phenomenon that the current dialogue state is closely related to the earlier dialogue states. In this paper, we try to introduce the slot embedding into the transformer to focus on those special tokens, which ever appear in earlier dialogue states. On the basis, we leverage the copy mechanism to predict the state over the dialogue utterances. Our model also imitates the architecture of reading comprehension model to make full use of the current utterances. The experimental results verify the benefits of the slot embedding, and our model achieves significant improvements than baselines on MultiWOZ 2.0 and MultiWOZ 2.1 datasets.

Structure Altering Effects of a Multicomponent Reading Intervention: An Application of the Direct and Inferential Mediation (DIME) Model of Reading Comprehension in Upper Elementary Grades

We evaluate the direct and inferential mediation (DIME) model for reading comprehension with a sample of struggling readers in Grades 3 to 5 (N = 364) in the context of a large-scale randomized controlled trial (RCT) investigating two theoretically distinct reading interventions (text processing + foundational skills [n = 117] or text processing only [n = 120]) and a control condition (n = 127). We investigate whether the intervention affects not just reading comprehension levels, but also how variables within the reading system interrelate. This approach allows the focus to shift from intervention as influencing a change in reading comprehension status to a complex set of processes. We fit structural equation models (SEMs) to evaluate the DIME model at baseline and a change model that included reading comprehension and word reading at posttest. There were no significant mean differences between groups in reading comprehension. However, significant differences emerged on the direct and indirect effects of background knowledge, vocabulary, word reading, strategies, and inferencing on comprehension across grade levels and treatment conditions. Related to treatment groups, background knowledge, vocabulary, and inferencing were significantly related to comprehension at posttest for students who received text processing and/or foundational skills interventions. The results have implications for the direct instruction of higher-order reading skills in the context of multicomponent interventions.

Individual reading styles: a narrative approach to understanding reading behaviour

This article reports on the findings from a study that intended to consider the validity of a model of reading comprehension. In the process of coding the resulting verbal protocols, the ‘feel’ or experience of the actual readers was lost, as was the individuality of the participants. Narratives were then developed to describe the experience of the reading and to recapture that individual nature of each reader’s approach to making sense of the texts they read. After those narratives were developed, some patterns emerged that are discussed in this article as reading styles. Some of these styles resulted in the reader developing a good understanding of what they were reading and others did not. The successful styles were classified in two ways, by the main strategy used and by the deliberateness of the readers’ construction of understanding. There are two pairs of successful reading styles discussed here, making four successful styles in total. Readers using a visualisation strategy are described as either “movie makers” for those who consciously constructed their mental movies, and as “Movie watchers” for those who experienced the process as being more receptive. The other strategy used in successful reading was questioning, either as an active “Hunter-seeker” approach or in the less-active “Wondering” style. Three styles that were not associated with successful reading are also described. Some implications for future research, and for teaching and learning are also discussed.

MTQA: Text-Based Multitype Question and Answer Reading Comprehension Model.

Text-based multitype question answering is one of the research hotspots in the field of reading comprehension models. Multitype reading comprehension models have the characteristics of shorter time to propose, complex components of relevant corpus, and greater difficulty in model construction. There are relatively few research works in this field. Therefore, it is urgent to improve the model performance. In this paper, a text-based multitype question and answer reading comprehension model (MTQA) is proposed. The model is based on a multilayer transformer encoding and decoding structure. In the decoding structure, the headers of the answer type prediction decoding, fragment decoding, arithmetic decoding, counting decoding, and negation are added for the characteristics of multiple types of corpora. Meanwhile, high-performance ELECTRA checkpoints are employed, and secondary pretraining based on these checkpoints and an absolute loss function are designed to improve the model performance. The experimental results show that the performance of the proposed model on the DROP and QUOREF corpora is better than the best results of the current existing models, which proves that the proposed MTQA model has high feature extraction and relatively strong generalization capabilities.

The Schematic Information-Processing (SIP) model of reading comprehension: Theoretical support for the utilization of text-relevant video segments to teach culturally unfamiliar texts in second/foreign language classrooms

This paper proposes a new theoretical framework for reading instruction in second/foreign language classrooms. The theoretical framework introduced in this paper is called the Schematic Information...

Analysis of English Multitext Reading Comprehension Model Based on Deep Belief Neural Network.

In order to solve the problems of low accuracy and low efficiency of answer prediction in machine reading comprehension, a multitext English reading comprehension model based on the deep belief neural network is proposed. Firstly, the paragraph selector in the multitext reading comprehension model is constructed. Secondly, the text reader is designed, and the deep belief neural network is introduced to predict the question answering probability. Finally, the popular English dataset of SQuAD is used for test analysis. The final results show that, after the comparative analysis of different learning methods, it is found that the English multitext reading comprehension model has a strong reading comprehension ability. In addition, two evaluation methods are used to score the overall performance of the model, which shows that the overall score of the English multitext reading comprehension model based on the deep confidence neural network is more than 90, and the efficiency will not be reduced because of the change of the number of documents in the dataset. The above results show that the use of the deep belief neural network to improve the probability generation performance of the model can well solve the task of English multitext reading comprehension, effectively reduce the difficulty of machine reading comprehension in multitask reading, and has a good guiding significance for promoting human convenient Internet knowledge acquisition.

Enhancement of Target‐Oriented Opinion Words Extraction with Multiview‐Trained Machine Reading Comprehension Model

Target-oriented opinion words extraction (TOWE) seeks to identify opinion expressions oriented to a specific target, and it is a crucial step toward fine-grained opinion mining. Recent neural networks have achieved significant success in this task by building target-aware representations. However, there are still two limitations of these methods that hinder the progress of TOWE. Mainstream approaches typically utilize position indicators to mark the given target, which is a naive strategy and lacks task-specific semantic meaning. Meanwhile, the annotated target-opinion pairs contain rich latent structural knowledge from multiple perspectives, but existing methods only exploit the TOWE view. To tackle these issues, we formulate the TOWE task as a question answering (QA) problem and leverage a machine reading comprehension (MRC) model trained with a multiview paradigm to extract targeted opinions. Specifically, we introduce a template-based pseudo-question generation method and utilize deep attention interaction to build target-aware context representations and extract related opinion words. To take advantage of latent structural correlations, we further cast the opinion-target structure into three distinct yet correlated views and leverage meta-learning to aggregate common knowledge among them to enhance the TOWE task. We evaluate the proposed model on four benchmark datasets, and our method achieves new state-of-the-art results. Extensional experiments have shown that the pipeline method with our approach could surpass existing opinion pair extraction models, including joint methods that are usually believed to work better.

QANet-based candidate answer rethink model for machine reading comprehension

The current model applied to the span extraction reading comprehension task fuses the information of context and question, and outputs the index with the highest probability calculated in the context as the prediction span. In this process, the model discards all the remaining candidate answers, which results in a waste of the available information in the candidate answers. Our model is designed to simulate the behaviour of human beings choosing multiple candidate answers and comprehensively judging the final answer in reading comprehension tasks. We propose the QANet-based candidate answer rethink model. The model interacts and fuses multiple candidate answers with context and question, prompting the model to obtain a more accurate answer by synthesising these three aspects of information. Experiments show that our model has made new progress in performance.

QANet-based candidate answer rethink model for machine reading comprehension

The current model applied to the span extraction reading comprehension task fuses the information of context and question, and outputs the index with the highest probability calculated in the context as the prediction span. In this process, the model discards all the remaining candidate answers, which results in a waste of the available information in the candidate answers. Our model is designed to simulate the behaviour of human beings choosing multiple candidate answers and comprehensively judging the final answer in reading comprehension tasks. We propose the QANet-based candidate answer rethink model. The model interacts and fuses multiple candidate answers with context and question, prompting the model to obtain a more accurate answer by synthesising these three aspects of information. Experiments show that our model has made new progress in performance.

S&I Reader: Multi-Granularity Gated Multi-Hop Skimming and Intensive Reading Model for Machine Reading Comprehension

Machine reading comprehension is a very challenging task, which aims to determine the answer span based on the given context and question. The newly developed pre-training language model has achieved a series of successes in various natural language understanding tasks with its powerful contextual representation ability. However, these pre-training language models generally lack the downstream processing structure for specific tasks, which limits further performance improvement. In order to solve this problem and deepen the model’s understanding of the question and context, this paper proposes S&I Reader. On the basis of the pre-training model, skimming, intensive reading, and gated mechanism modules are added to simulate the behavior of humans reading text and filtering information. Based on the idea of granular computing, a multi-granularity module for computing context granularity and sequence granularity is added to the model to simulate the behavior of human beings to understand the text from words to sentences, from parts to the whole. Compared with the previous machine reading comprehension model, our model structure is novel. The skimming module and multi-granularity module proposed in this paper are used to solve the problem that the previous model ignores the key information of the text and cannot understand the text with multi granularity. Experiments show that the model proposed in this paper is effective for both Chinese and English datasets. It can better understand the question and context and give a more accurate answer. The performance has made new progress on the basis of the baseline model.

Extensive and Intensive: A BERT-based machine reading comprehension model with two reading strategies

Enabling machines to read, process, and understand natural language documents is a coveted goal of artificial intelligence. However, this task is extremely challenging, and most existing models lack the ability to perform complex reasoning. Considering that humans often read documents roughly first when understanding a problem, this paper proposes a new model that attempts to mimic the reasoning process of human readers. Our model performs a extensive read and a intensive read of the document separately, and then combines the information obtained from both reading methods to finally find a satisfactory answer. Finally, by experimenting within RACE dataset and comparing with the baseline model BERT, the feasibility and effectiveness of our proposed model can be illustrated.

Beat the AI: Investigating Adversarial Human Annotation for Reading Comprehension

Innovations in annotation methodology have been a catalyst for Reading Comprehension (RC) datasets and models. One recent trend to challenge current RC models is to involve a model in the annotation process: Humans create questions adversarially, such that the model fails to answer them correctly. In this work we investigate this annotation methodology and apply it in three different settings, collecting a total of 36,000 samples with progressively stronger models in the annotation loop. This allows us to explore questions such as the reproducibility of the adversarial effect, transfer from data collected with varying model-in-the-loop strengths, and generalization to data collected without a model. We find that training on adversarially collected samples leads to strong generalization to non-adversarially collected datasets, yet with progressive performance deterioration with increasingly stronger models-in-the-loop. Furthermore, we find that stronger models can still learn from datasets collected with substantially weaker models-in-the-loop. When trained on data collected with a BiDAF model in the loop, RoBERTa achieves 39.9F1 on questions that it cannot answer when trained on SQuAD—only marginally lower than when trained on data collected using RoBERTa itself (41.0F1).

Broadening the Lens on the Science of Reading: A Multifaceted Perspective on the Role of Academic Language in Text Understanding

ABSTRACTMerging evidence from psychological models of reading comprehension, ethnographic research on language and literacy, and textual linguistics lines of research, the authors take the position that psychological models of reading comprehension often overlook written language comprehension and production as context‐embedded, sociocultural processes. Using the example of academic language comprehension by middle grade learners, the authors present a series of studies that have informed the science of reading by making visible a precise set of high‐utility academic language skills that support informational text comprehension during middle childhood (ages 9–14). Studies have suggested that these skills develop gradually throughout adolescence and significantly contribute to reading comprehension. Drawing on this research with the goal of informing the science of reading, the authors suggest that academic language comprehension involves for the reader (a) familiarity with a set of academic language forms commonly found in school texts, (b) experience with the sociocultural practices of understanding and using the academic language of text within a particular sociocultural community, and (c) aligning with or resisting the reader identities implied by the language of a text. The authors suggest that innovative research and pedagogical approaches are needed to broaden the conceptualization of language and reading comprehension relations from purely cognitive into one that embraces the reader’s interaction with a text as a sociocultural phenomenon.

Research and Design of Chinese Question Answer System for Heart Disease

This article is aimed at the current common users to obtain information on heart disease online, citing the current popular question and answer system. this article introduces a reading comprehension model to find answers that match the questions. In order to make the reading comprehension model have better application in the question answering system, the following work is done in this article: (1) Quoted word vectors are used to process Chinese text. Since Chinese is more complicated than English, quoted word vectors are currently a better method for processing Chinese text. (2) Construct a question answering system, which can be used to get answers to questions in a paragraph.

Examining the L2 reading comprehension ability of adult ELLs: Developing a diagnostic test within the cognitive diagnostic assessment framework

Cognitive diagnostic assessment (CDA) has been applied to language assessment in a number of studies in which a diagnostic classification model (DCM) was retrofitted to the results of a non-diagnostic assessment. However, the need to apply CDA through utilization of an inductive rather than a retrofitted approach has been a recurrent theme in these studies. Thus, this study aimed to develop a diagnostic L2 reading comprehension test in English to investigate adult examinees’ reading performances in an EFL (English as a Foreign Language) academic setting. The test was based on a cognitive model of L2 reading comprehension and was administered to a sample of 1058 examinees across Turkey. The results were analyzed using log-linear cognitive diagnosis modeling (LCDM), which is one of the general DCM families subsuming other core DCMs. The findings of the study indicated that obtaining fine-grained diagnostic information about examinees’ performances in a given domain would be possible by coupling an adequate understanding of the construct with a CDA framework.

Multi-task reading for intelligent legal services

Since legal data contains both structured data and unstructured data, it is a great challenge to implement machine reading comprehension technology in empirical analysis of law. This paper proposes a multi-tasking reading for intelligent legal services, which applies statistical analysis and machine reading comprehension techniques, and can process both structured and unstructured data. At the same time, this paper proposes a machine reading comprehension model that can perform multi-task learning, LegalSelfReader, which can solve the problem of diversity of questions. In the experiment of the legal reading comprehension dataset CJRC, the model proposed in this paper is far superior to the two classic models of BIDAF and Bert in three evaluation indicators. And our model is also better than some models published by HFL(Harbin Institute of Technology and iFly Joint Lab), and has also achieved lower consumption in training costs. At the same time, in the experiment of visualizing the attention value, it also demonstrates that the model proposed in this paper has a stronger ability to extract evidence.

Language transfer theory and its policy implications: exploring interdependence between Luganda, Runyankole-Rukiga, and English in Uganda

ABSTRACT The aim of the current study is to expand our understanding of cross-linguistic transfer (CLT) to new languages and settings under-represented in the research literature. Empirical evidence of CLT is almost exclusively demonstrated in studies examining European and Asian languages in high-income countries. Literacy data were collected from 3561 second-grade students in one of two languages in Uganda – Luganda and Runyankole-Rukiga – and English. Structural equation modelling was used to test a latent construct of alphabetical knowledge and a path model of bilingual reading comprehension. Findings suggest that second-grade students transfer decoding and reading comprehension skills from their first to second language, and initial evidence that the interdependence varies according to the language used in the classroom. Implications for language-in-education policies are discussed.

A Conceptual Framework: The Integration of SQ3R in Flipped Classroom Model for Chinese Reading Comprehension

Reading comprehension is one of the important elements in reading skills. Through reading comprehension, students will be able to know the meaning and gain the main information from the reading materials. An effective reading strategy will be a better way to help students understand more about reading. Meanwhile, internet technology is widely used in the education field to promote learning that occurs in the flipped model. This study presents a practical teaching strategy to integrate the SQ3R reading strategy with the help of technology under a new teaching model. Literature reviews on reading comprehension, SQ3R strategy and flipped classroom model are also presented in this paper. This study will employ quantitative approach using surveys technique. Total of 118 year four students from three different primary school in Klang district will be involved in this survey research. A questionnaire and structured interview will be carried out to examine the students’ perception toward the integration of SQ3R in Flipped Classroom Model for Chinese reading comprehension. The primary aim of this study is to promote the Chinese reading comprehension learning process using SQ3R strategy in the flipped classroom teaching model. Through the proposed framework in this study, the SQ3R strategy in the flipped classroom model can be applied in the pedagogical practice. Therefore, this study should be valuable to researchers to present future studies and teachers who wish to improve their teaching methods.

Improving the robustness of machine reading comprehension model with hierarchical knowledge and auxiliary unanswerability prediction

Machine Reading Comprehension (MRC) aims to understand a passage and answer a series of related questions. With the development of deep learning and the release of large-scale MRC datasets, many end-to-end MRC neural networks have achieved remarkable success. However, these models are fragile and lack of robustness when there are some imperceptible adversarial perturbations in the input. In this paper, we propose an MRC model which has two main components to improve the robustness. On the one hand, we enhance the representation of the model by leveraging hierarchical knowledge from external knowledge bases. On the other hand, we introduce an auxiliary unanswerability prediction module and perform supervised multi-task learning along with a span prediction task. Experimental results on benchmark datasets show that our model can achieve consistent improvement compared with other strong baselines.

‘I’m probably just gonna skim’: an assessment of undergraduate students’ primary scientific literature reading approaches

ABSTRACT We investigated undergraduate students’ approaches to reading primary scientific literature (PSL). Self-report surveys and think-aloud reading interviews were used to uncover students’ approaches to PSL with respect to evidence finding, prioritisation of paper sections, and reading skill in relation to task context. Self-report and observational interview data were also analysed to investigate the relationship between students’ self-reported and actual reading approaches. Our findings indicate that undergraduate students exhibit a spectrum of approaches to evidence finding, that many use pre-interpreted text to make meaning of PSL, and that student reading approaches are enhanced by task context. Data also indicate a misalignment between students’ self-reported and actual reading approaches, with students generally over-confident in their abilities. We interpret these findings using deep vs. surface learning criteria and the Structure Building Model (SBM) of reading comprehension. Our exploratory research extends the small body of literature on student PSL reading and indicates that more in-depth work needs to be done to support the development of educational interventions aimed at enhancing student reading approaches.