Answering System Research Articles

Unknown-class recognition adversarial network for open set domain adaptation fault diagnosis of rotating machinery

Unknown-class recognition adversarial network for open set domain adaptation fault diagnosis of rotating machinery

A Graph-Based Keyword Extraction Method for Academic Literature Knowledge Graph Construction

In this paper, we construct an academic literature knowledge graph based on the relationship between documents to facilitate the storage and research of academic literature data. Keywords are an important type of node in the knowledge graph. To solve the problem that there are no keywords in some documents for several reasons in the process of knowledge graph construction, an improved keyword extraction algorithm called TP-CoGlo-TextRank is proposed by using word frequency, position, word co-occurrence frequency, and a word embedding model. By combining the word frequency and position in the document, the importance of words is distinguished. By introducing the GloVe word-embedding model, which brings the external knowledge of documents into the TextRank algorithm, and combining the internal word co-occurrence frequency in the documents, the word-adjacency relationship is transferred non-uniformly. Finally, the words with the highest scores are combined into phrases if they are adjacent in the original text. The validity of the TP-CoGlo-TextRank algorithm is verified by experiments. On this basis, the Neo4j graph database is used to store and display the academic literature knowledge graph, to provide data support for research tasks such as text clustering, automatic summarization, and question-answering systems.

NEUROSYNTH: Enhancing Cognitive Computing using Deep Neural Networks

This study presents an innovative approach to enhance question-answering (QA) systems that utilize a RoBERTa-based architecture and complexity-enhanced input features. The work is divided into four primary parts: training methodology, feature engineering, building models, and data preprocessing. We propose a Python function that uses readability measures and natural language processing techniques to calculate the linguistic difficulty metrics for input sentences. TensorFlow Datasets (TFDS) are then used to load and preprocess the SquAD (Stanford Question Answering Dataset) dataset to enable effective training. Word embeddings from previously trained GloVe vectors are integrated with complexity metrics to prepare input features that add contextual information to the input representation. With the inclusion of the enriched features, a distinctive question-answering model based on the RoBERTa architecture is trained using the AdamW optimizer and CrossEntropyLoss. Iterative epochs are used in the training process to optimize the model's parameters and minimize the loss function. An independent validation dataset is used to evaluate the model's performance, proving the usefulness of the suggested method in improving the accuracy and robustness of the QA system. All in all, this work offers an organized strategy for improving the quality of systems by fusing cutting-edge neural architecture with input properties that are increased by complexity.

Large Language Model-Based Evaluation of Medical Question Answering Systems: Algorithm Development and Case Study.

Healthcare systems are increasingly resource constrained, leaving less time for important patient-provider interactions. Conversational agents (CAs) could be used to support the provision of information and to answer patients' questions. However, information must be accessible to a variety of patient populations, which requires understanding questions expressed at different language levels. This study describes the use of Large Language Models (LLMs) to evaluate predefined medical content in CAs across patient populations. These simulated populations are characterized by a range of health literacy. The evaluation framework includes both fully automated and semi-automated procedures to assess the performance of a CA. A case study in the domain of mammography shows that LLMs can simulate questions from different patient populations. However, the accuracy of the answers provided varies depending on the level of health literacy. Our scalable evaluation framework enables the simulation of patient populations with different health literacy levels and helps to evaluate domain specific CAs, thus promoting their integration into clinical practice. Future research aims to extend the framework to CAs without predefined content and to apply LLMs to adapt medical information to the specific (health) literacy level of the user.

Ecological aspects and relationships of the emblematic Vachellia spp. exposed to anthropic pressures and parasitism in natural hyper-arid ecosystems: ethnobotanical elements, morphology, and biological nitrogen fixation

Main conclusionEmblematic Vachellia spp. naturally exposed to hyper-arid conditions, intensive grazing, and parasitism maintain a high nitrogen content and functional mutualistic nitrogen-fixing symbioses.AlUla region in Saudi Arabia has a rich history regarding mankind, local wildlife, and fertility islands suitable for leguminous species, such as the emblematic Vachellia spp. desert trees. In this region, we investigated the characteristics of desert legumes in two nature reserves (Sharaan and Madakhil), at one archaeological site (Hegra), and in open public domains et al. Ward and Jabal Abu Oud. Biological nitrogen fixation (BNF), isotopes, and N and C contents were investigated through multiple lenses, including parasitism, plant tissues, species identification, plant maturity, health status, and plant growth. The average BNF rates of 19 Vachellia gerrardii and 21 Vachellia tortilis trees were respectively 39 and 67%, with low signs of inner N content fluctuations (2.10–2.63% N) compared to other co-occurring plants. The BNF of 23 R. raetam was just as high, with an average of 65% and steady inner N contents of 2.25 ± 0.30%. Regarding parasitism, infected Vachellia trees were unfazed compared to uninfected trees, thereby challenging the commonly accepted detrimental role of parasites. Overall, these results suggest that Vachellia trees and R. raetam shrubs exploit BNF in hyper-arid environments to maintain a high N content when exposed to parasitism and grazing. These findings underline the pivotal role of plant-bacteria mutualistic symbioses in desert environments. All ecological traits and relationships mentioned are further arguments in favor of these legumes serving as keystone species for ecological restoration and agro-silvo-pastoralism in the AlUla region.Graphical abstract

CLOSED-DOMAIN QUESTION-ANSWERING SYSTEM FOR INDIAN LEGAL ACTS

: In Today’s digital age, access to legal knowledge among students is crucial for fostering a well-informed citizenry. However, comprehending the intricacies of provisions of Indian Legal Acts, especially in a complex legal system like India’s, can be daunting, even for seasoned professionals, let alone for educational purposes among children. To bridge this gap and foster legal literacy from a young age, we introduced a PDF-based Closed-domain Question-Answering (CDQA) System for Indian Legal Acts. Our system simplifies educating students about provisions of Indian Legal Acts such as RTI(Right to Education), RTE(Right to Education), Anti-Dowry Act, etc. By leveraging natural language processing(NLP) techniques and machine learning algorithms, our system, powered by LangChain, enables users to pose questions in natural language related to Indian Legal Acts. LangChain represents a groundbreaking advancement in question-answer systems, harnessing the power of cutting-edge language models to provide accurate and comprehensive responses to user queries. Developed based on state-of-the-art NLP techniques, LangChain is a versatile and highly adaptable tool that offers a versatile solution with transformative potential across diverse industries, including education, healthcare, legal research, customer support, etc.

REAL-TIME SPEECH TO TEXT CONVERSION & ANSWER EVALUATION

To improve user engagement, a new question-answering system including speech recognition technology is presented in this study. Users react verbally to queries presented to them via a command-line interface built into the system. Advanced speech-to-text conversion techniques are used to convert these spoken responses into text for analysis. The system compares user responses with predefined proper responses using cosine similarity to assess input correctness. An SMTP server is used by the system to automatically send quiz scores to users via email after the quiz is finished. Showcasing the merging of spoken conversation and automated evaluation in educational environments, this ground-breaking solution offers a simplified way to interactive learning and assessment. Keywords—Question-answering, Speech recognition, Command-line interface, Speech-to-text conversion, Cosine similarity [8], User input accuracy, Quiz scores, Email notification, SMTP [4] server, Interactive learning.

GPB and BAC: two novel models towards building an intelligent motor fault maintenance question answering system

Generally, the existing methods for constructing a knowledge graph used in a question answering system adopted two different models respectively, one is for identifying entities, and the other is for extracting relationships between entities. However, this method may reduce the quality of knowledge because it is very difficult to keep contextual information consistent with the same entities in the two different models. To address this issue, this paper proposes a model called GPB (GlobalPointer + BiLSTM) which integrates the BiLSTM into GlobalPointer through concatenation operations to simultaneously guarantee the rationality of identified entities and relationships between entities. In addition, to enhance the user experience using an intelligent motor fault maintenance question answering system, a model called BAC (BiLSTM + Attention + CRF) is proposed to identify named entities in user questions, and the BERT-wwm model is used to classify user intentions to improve the quality of answers. Finally, to verify the advantages of the proposed model GPB and BAC, comparative experiments and real application effects of the developed question answering system are demonstrated on our built motor fault maintenance dataset. The experimental results indicate that the constructed knowledge graph and developed question answering system provide engineers with high-quality motor maintenance knowledge services.

Interactive Question Answering Systems: Literature Review

Question-answering systems are recognized as popular and frequently effective means of information seeking on the web. In such systems, information seekers can receive a concise response to their queries by presenting their questions in natural language. Interactive question answering is a recently proposed and increasingly popular solution that resides at the intersection of question answering and dialogue systems . On the one hand, the user can ask questions in normal language and locate the actual response to her inquiry; on the other hand, the system can prolong the question-answering session into a dialogue if there are multiple probable replies, very few, or ambiguities in the initial request. By permitting the user to ask more questions, interactive question answering enables users to interact with the system and receive more precise results dynamically. This survey offers a detailed overview of the interactive question-answering methods that are prevalent in current literature. It begins by explaining the foundational principles of question-answering systems, hence defining new notations and taxonomies to combine all identified works inside a unified framework. The reviewed published work on interactive question-answering systems is then presented and examined in terms of its proposed methodology, evaluation approaches, and dataset/application domain. We also describe trends surrounding specific tasks and issues raised by the community, so shedding light on the future interests of scholars. Our work is further supported by a GitHub page synthesising all the major topics covered in this literature study. https://sisinflab.github.io/interactive-question-answering-systems-survey/

Knowledge Graph Multi-Hop Question Answering Based on Dependent Syntactic Semantic Augmented Graph Networks

In the rapidly evolving domain of question answering systems, the ability to integrate machine comprehension with relational reasoning stands paramount. This paper introduces a novel architecture, the Dependent Syntactic Semantic Augmented Graph Network (DSSAGN), designed to address the intricate challenges of multi-hop question answering. By ingeniously leveraging the synergy between syntactic structures and semantic relationships within knowledge graphs, DSSAGN offers a breakthrough in interpretability, scalability, and accuracy. Unlike previous models that either fall short in handling complex relational paths or lack transparency in reasoning, our framework excels by embedding a sophisticated mechanism that meticulously models multi-hop relations and dynamically prioritizes the syntactic–semantic context.

Nanjing Yunjin intelligent question-answering system based on knowledge graphs and retrieval augmented generation technology

Nanjing Yunjin, a traditional Chinese silk weaving craft, is celebrated globally for its unique local characteristics and exquisite workmanship, forming an integral part of the world's intangible cultural heritage. However, with the advancement of information technology, the experiential knowledge of the Nanjing Yunjin production process is predominantly stored in text format. As a highly specialized and vertical domain, this information is not readily convert into usable data. Previous studies on a knowledge graph-based Nanjing Yunjin Question-Answering System have partially addressed this issue. However, knowledge graphs need to be constantly updated and rely on predefined entities and relationship types. Faced with ambiguous or complex natural language problems, knowledge graph information retrieval faces some challenges. Therefore, this study proposes a Nanjing Yunjin Question-Answering System that integrates Knowledge Graphs and Retrieval Augmented Generation techniques. In this system, the ROBERTA model is first utilized to vectorize Nanjing Yunjin textual information, delving deep into textual semantics to unveil its profound cultural connotations. Additionally, the FAISS vector database is employed for efficient storage and retrieval of Nanjing Yunjin information, achieving a deep semantic match between questions and answers. Ultimately, related retrieval results are fed into the Large Language Model for enhanced generation, aiming for more accurate text generation outcomes and improving the interpretability and logic of the Question-Answering System. This research merges technologies like text embedding, vectorized retrieval, and natural language generation, aiming to overcome the limitations of knowledge graphs-based Question-Answering System in terms of graph updating, dependency on predefined types, and semantic understanding. System implementation and testing have shown that the Nanjing Yunjin Intelligent Question-Answering System, constructed on the basis of Knowledge Graphs and Retrieval Augmented Generation, possesses a broader knowledge base that considers context, resolving issues of polysemy, vague language, and sentence ambiguity, and efficiently and accurately generates answers to natural language queries. This significantly facilitates the retrieval and utilization of Yunjin knowledge, providing a paradigm for constructing Question-Answering System for other intangible cultural heritages, and holds substantial theoretical and practical significance for the deep exploration and discovery of the knowledge structure of human intangible heritage, promoting cultural inheritance and protection.

Retrieve What You Need: A Mutual Learning Framework for Open-domain Question Answering

Abstract An open-domain question answering (QA) system usually follows a retrieve-then-read paradigm, in which a retriever is used to retrieve relevant passages from a large corpus, and then a reader generates answers based on the retrieved passages and the original question. In this paper, we propose a simple and novel mutual learning framework to improve the performance of retrieve-then-read-style models via an intermediate module named the knowledge selector, which we train with reinforcement learning. The key benefits of our proposed intermediate module are: 1) no requirement for additional annotated question-passage pairs; 2) improvements in both retrieval and QA performance, as well as computational efficiency, compared to prior competitive retrieve-then-read models; 3) with no finetuning, improvement in the zero-shot performance of large-scale pre-trained language models, e.g., ChatGPT, by encapsulating the input with relevant knowledge without violating the input length constraint.

ChatBBNJ: a question–answering system for acquiring knowledge on biodiversity beyond national jurisdiction

The marine biodiversity in Areas beyond national jurisdiction (ABNJ), encompassing approximately two-thirds of the global ocean, is persistently declining. In 2023, the agreement on the Conservation and Sustainable Use of Marine Biodiversity of Areas Beyond National Jurisdiction (BBNJ) was officially adopted. Implementing the BBNJ Agreement has the potential to effectively meet global needs for preserving marine biodiversity. Nevertheless, the implementation requires dealing with thousands of legal clauses, and the parties participating in the process lack adequate means to acquire knowledge connected to BBNJ. This paper introduces ChatBBNJ, a highly efficient question-answering system that combines a novel data engineering technique with large language models (LLMs) of Natural Language Processing (NLP). The system aims to efficiently provide stakeholders with BBNJ-related knowledge, thereby facilitating and enhancing their comprehension and involvement with the subject matter. The experimental results demonstrate that the proposed ChatBBNJ exhibits superior expertise in the BBNJ domain, outperforming baseline models in terms of precision, recall, and F1-scores. The successful deployment of the suggested system is expected to greatly assist stakeholders in acquiring BBNJ knowledge and facilitating the effective implementation of the BBNJ Agreement. Therefore, this is expected to contribute to the conservation and sustainable use of marine biodiversity in ABNJ.

A Review on Conversational Question Answering (CQA)

The internet is constantly changing how people communicate with one another, manage their daily lives, and share the information. As a result of digitization, automated Question Answering Systems are becoming more important in extracting useful information from the knowledge sources. In general, Question Answering (QA)mechanisms are intended to answer a specific question only once (so-called single-turn).However, to satisfy a user's information needs, the Conversational Question Answering (CQA) system must grasp the given context and engage in multi-turn QA. While single-turn QA leads the majority of current research, multi-turn QA has rapidly gained prominence due to the availability of large-scale multi-turn QA datasets and the development of “pre-trained language models”. In this research paper, we have presented a comprehensive review on the Conversational Question Answer (CQA) field which includes need of CQA followed by categorization of QA system, list of available datasets and various evaluation metrics.

A visual question and answering system with support for compound emotions using facial landmark identification with MediaPipe and CNN classifier

Visual Question and Answering is a fast-evolving field of research where we attempt to answer the question based on an image as a context. In this paper, we try to add support for answering questions that are based on emotions with face detection and basic emotion extraction. With the basic emotions, we further try to find out compound emotions to expand the area of answerable questions. We complete the said operations in two stages: i) Caption generation. ii) Question and answering with that caption as context. The context generation process can be further divided into i) general caption generator ii) emotional caption generator. In general caption generator, we try to find a single sentence providing a basic information about the input context image, using an Attention [9] based model. In the later phase, the emotional caption generator finds the information of the emotions of the faces of the image using RetinaFace [20] and DeepFace [1]. In emotional caption generator, apart from basic emotion generation, we also try to find compound emotions depicted by each of the faces by learning the patterns in facial landmarks generated by MediaPipe [7] with CNN based architecture. In this module, we also generate a set of know question and answers to use in the next phase. As for the question and answering part, we first use BERT [4] for finding an answer, given a question and the captions generated as context of the question. If the confidence of answer exceeds a threshold value, we report that answer as final. If not, we try to find the known question similar to the asked question. The answer generated then is the final answer.

Question-answering system extracts information on injection drug use from clinical notes

BackgroundInjection drug use (IDU) can increase mortality and morbidity. Therefore, identifying IDU early and initiating harm reduction interventions can benefit individuals at risk. However, extracting IDU behaviors from patients’ electronic health records (EHR) is difficult because there is no other structured data available, such as International Classification of Disease (ICD) codes, and IDU is most often documented in unstructured free-text clinical notes. Although natural language processing can efficiently extract this information from unstructured data, there are no validated tools.MethodsTo address this gap in clinical information, we design a question-answering (QA) framework to extract information on IDU from clinical notes for use in clinical operations. Our framework involves two main steps: (1) generating a gold-standard QA dataset and (2) developing and testing the QA model. We use 2323 clinical notes of 1145 patients curated from the US Department of Veterans Affairs (VA) Corporate Data Warehouse to construct the gold-standard dataset for developing and evaluating the QA model. We also demonstrate the QA model’s ability to extract IDU-related information from temporally out-of-distribution data.ResultsHere, we show that for a strict match between gold-standard and predicted answers, the QA model achieves a 51.65% F1 score. For a relaxed match between the gold-standard and predicted answers, the QA model obtains a 78.03% F1 score, along with 85.38% Precision and 79.02% Recall scores. Moreover, the QA model demonstrates consistent performance when subjected to temporally out-of-distribution data.ConclusionsOur study introduces a QA framework designed to extract IDU information from clinical notes, aiming to enhance the accurate and efficient detection of people who inject drugs, extract relevant information, and ultimately facilitate informed patient care.

Multilingual Question Answering for Malaysia History with Transformer-based Language Model

In natural language processing (NLP), a Question Answering System (QAS) refers to a system or model that is designed to understand and respond to user queries in natural language. As we navigate through the recent advancements in QAS, it can be observed that there is a paradigm shift of the methods used from traditional machine learning and deep learning approaches towards transformer-based language models. While significant progress has been made, the utilization of these models for historical QAS and the development of QAS for Malay language remain largely unexplored. This research aims to bridge the gaps, focusing on developing a Multilingual QAS for history of Malaysia by utilizing a transformer-based language model. The system development process encompasses various stages, including data collection, knowledge representation, data loading and pre-processing, document indexing and storing, and the establishment of a querying pipeline with the retriever and reader. A dataset with a collection of 100 articles, including web blogs related to the history of Malaysia, has been constructed, serving as the knowledge base for the proposed QAS. A significant aspect of this research is the use of the translated dataset in English instead of the raw dataset in Malay. This decision was made to leverage the effectiveness of well-established retriever and reader models that were trained on English data. Moreover, an evaluation dataset comprising 100 question-answer pairs has been created to evaluate the performance of the models. A comparative analysis of six different transformer-based language models, namely DeBERTaV3, BERT, ALBERT, ELECTRA, MiniLM, and RoBERTa, has been conducted, where the effectiveness of the models was examined through a series of experiments to determine the best reader model for the proposed QAS. The experimental results reveal that the proposed QAS achieved the best performance when employing RoBERTa as the reader model. Finally, the proposed QAS was deployed on Discord and equipped with multilingual support through the incorporation of language detection and translation modules, enabling it to handle queries in both Malay and English. Doi: 10.28991/ESJ-2024-08-02-019 Full Text: PDF

Hydrodynamic performance of vertical cylindrical wave energy absorbers in front of a vertical wall

This paper studies the hydrodynamic performance of vertical cylindrical absorbers in front of a vertical wall. All the absorbers are independent of each other and restricted to only the heave motion. Based on a linear potential flow theory, an analytical solution is developed for the problems of wave diffraction and radiation by absorbers. In the solving procedure, the hydrodynamic problem is first transformed into an equivalent problem in an open water domain using the image principle. The number of absorbers in the equivalent problem is twice that in the real problem, and the plane layout is symmetric about the original vertical wall. The velocity potential of the fluid domain is obtained using the method of variable separation, and the unknown expansion coefficients in the velocity potential are determined by the matched boundary conditions. The heave excitation force, added mass, radiation damping, motion response, and energy capture width of the absorbers are calculated. Case studies are presented to show the effects of the wall reflection and hydrodynamic interaction on the energy extraction performance of the wave energy converter (WEC) system. Subsequently, the WEC performance under the action of irregular waves is analyzed by considering an incident wave spectrum, and the mean annual absorbed power of the device is estimated by considering the wave data statistics at the actual sites. The results indicate that when the wave motion resonates with the absorber motion, the energy extraction performance of the absorbers is significantly improved. The performance of the absorbers can be effectively improved when the structures are close to the antinodes of a standing wave field. By designing a reasonable plane layout, the hydrodynamic interaction can play a constructive role in the performance of the WEC system.

Weight Averaging and re-adjustment ensemble for QRCD

Question Answering (QA) is a prominent task in the field of Natural Language Processing (NLP) with extensive applications. Recently, there has been a notable surge in research interest concerning the development of QA systems for the Holy Qur’an, an Islamic religious text. The Qur’an Reading Comprehension Dataset (QRCD) Malhas and Elsayed (2020) is a highly commendable effort in this respect. It stands as the first benchmark dataset specifically designed for a set of directly answerable questions from the Qur’an. Each question in the dataset is meticulously labeled with all potential answers sourced from the Holy Qur’an. From our perspective, the main challenge in QRCD stems from the limited volume of training data it offers. As a solution we propose an innovative approach to build a Deep Neural Network (DNN) ensemble, centered around Ara-Electra model (Antoun et al., 2021), that we called Weight Averaging and Re-adjustment (WAR) model. The model is constructed by computing running averages of all model states that evolve during a single training session and ensuring that model weights are readjusted prior to each training epoch, in order to hold it back from over fitting the training data. The scheme results in a single standalone model that exhibits the benefits of multi-model ensembles. It is distinguished from other ensembles proposed for QRCD that accumulates outputs from multiple expert models and employs classic techniques like hard voting or score averaging on output probabilities to build unified results. Each expert model costs individual training time and compute resources. The WAR model outperforms existing systems with improved generalization over unseen data. It achieves F1, partial Reciprocal Rank (pRR), and exact-match (EM) scores of 0.567, 0.60 and 0.29 respectively, exceeding best reported QRCD scores by 3%, 1.5% and 0.69% respectively. Notably, we are comparing our results with the top scores from different models, highlighting our model’s consistent performance across all three metrics.

Application of Multimodal Transformer Model in Intelligent Agricultural Disease Detection and Question-Answering Systems.

In this study, an innovative approach based on multimodal data and the transformer model was proposed to address challenges in agricultural disease detection and question-answering systems. This method effectively integrates image, text, and sensor data, utilizing deep learning technologies to profoundly analyze and process complex agriculture-related issues. The study achieved technical breakthroughs and provides new perspectives and tools for the development of intelligent agriculture. In the task of agricultural disease detection, the proposed method demonstrated outstanding performance, achieving a precision, recall, and accuracy of 0.95, 0.92, and 0.94, respectively, significantly outperforming the other conventional deep learning models. These results indicate the method's effectiveness in identifying and accurately classifying various agricultural diseases, particularly excelling in handling subtle features and complex data. In the task of generating descriptive text from agricultural images, the method also exhibited impressive performance, with a precision, recall, and accuracy of 0.92, 0.88, and 0.91, respectively. This demonstrates that the method can not only deeply understand the content of agricultural images but also generate accurate and rich descriptive texts. The object detection experiment further validated the effectiveness of our approach, where the method achieved a precision, recall, and accuracy of 0.96, 0.91, and 0.94. This achievement highlights the method's capability for accurately locating and identifying agricultural targets, especially in complex environments. Overall, the approach in this study not only demonstrated exceptional performance in multiple tasks such as agricultural disease detection, image captioning, and object detection but also showcased the immense potential of multimodal data and deep learning technologies in the application of intelligent agriculture.