Search Results (14)

Civil aviation airworthiness regulations are the fundamental basis for the design and operational safety of aircraft. Their provisions exhibit a high degree of specialization, cross-disciplinary complexity, and hierarchical structure. Moreover, the regulations are frequently updated, posing unique challenges for automated question-answering systems. While large language models (LLMs) have demonstrated remarkable capabilities in dialog and reasoning; however, they still face challenges such as difficulties in knowledge updating and a scarcity of high-quality domain-specific datasets when tackling knowledge-intensive tasks in the field of civil aviation regulations. This study introduces a retrieval-augmented generation (RAG) approach that integrates retrieval modules with generative models to enable more efficient knowledge acquisition and updating, encompassing data processing and retrieval-based reasoning. The data processing stage comprises document conversion, information extraction, and document parsing modules. Additionally, a high-quality airworthiness regulation QA dataset was specifically constructed, covering multiple-choice, true/false, and fill-in-the-blank questions, with a total of 4688 entries. The retrieval-based reasoning stage employs vector search and re-ranking strategies, combined with prompt optimization, to enhance the model’s reasoning capabilities in specific airworthiness certification regulation comprehension tasks. A series of experiments demonstrate the effectiveness of the retrieval-augmented generation approach in this domain, significantly improving answer accuracy and retrieval hit rates. Full article

As large language models (LLMs) increasingly enter specialized domains, inference without external resources often leads to knowledge gaps, opaque reasoning, and hallucinations. To address these challenges in power equipment defect grading, we propose a zero-shot question-answering framework that requires no task-specific examples. Our system performs two-stage retrieval—first using a Sentence-BERT model fine-tuned on power equipment maintenance texts for coarse filtering, then combining TF-IDF and semantic re-ranking for fine-grained selection of the most relevant knowledge snippets. We embed both the user query and the retrieved evidence into a Chain-of-Thought (CoT) prompt, guiding the pre-trained LLM through multi-step reasoning with self-validation and without any model fine-tuning. Experimental results show that on a held-out test set of 218 inspection records, our method achieves a grading accuracy of 54.2%, which is 6.0 percentage points higher than the fine-tuned BERT baseline at 48.2%; an Explanation Coherence Score (ECS) of 4.2 compared to 3.1 for the baseline; a mean retrieval latency of 28.3 ms; and an average LLM inference time of 5.46 s. Ablation and sensitivity analyses demonstrate that a fine-stage retrieval pool size of k = 30 offers the optimal trade-off between accuracy and latency; human expert evaluation by six senior engineers yields average Usefulness and Trustworthiness scores of 4.1 and 4.3, respectively. Case studies across representative defect scenarios further highlight the system’s robust zero-shot performance. Full article

Knowledge graph completion (KGC) involves inferring missing entities or relationships within a knowledge graph, playing a crucial role across various domains, including intelligent question answering, recommendation systems, and dialogue systems. Traditional knowledge graph embedding (KGE) methods have proven effective in utilizing structured data and relationships. However, these methods often overlook the vast amounts of unstructured data and the complex reasoning capabilities required to handle ambiguous queries or rare entities. Recently, the rapid development of large language models (LLMs) has demonstrated exceptional potential in text comprehension and contextual reasoning, offering new prospects for KGC tasks. By using traditional KGE to capture the structural information of entities and relations to generate candidate entities and then reranking them with a generative LLM, the output of the LLM can be constrained to improve reliability. Despite this, new challenges, such as omissions and incorrect responses, arise during the ranking process. To address these issues, a knowledge-guided LLM reasoning for knowledge graph completion (KLR-KGC) framework is proposed. This model retrieves two types of knowledge from the knowledge graph—analogical knowledge and subgraph knowledge—to enhance the LLM’s logical reasoning ability for specific tasks while injecting relevant additional knowledge. By integrating a chain-of-thought (CoT) prompting strategy, the model guides the LLM to filter and rerank candidate entities, constraining its output to reduce omissions and incorrect responses. The framework aims to learn and uncover the latent correspondences between entities, guiding the LLM to make reasonable inferences based on supplementary knowledge for more accurate predictions. The experimental results demonstrate that on the FB15k-237 dataset, KLR-KGC outperformed the entity generation model (CompGCN), achieving a 4.8% improvement in MRR and a 5.8% improvement in Hits@1. Full article

Despite the remarkable progress of large language models (LLMs) in understanding and generating unstructured text, their application in structured data domains and their multi-role capabilities remain underexplored. In particular, utilizing LLMs to perform complex reasoning tasks on knowledge graphs (KGs) is still an emerging area with limited research. To address this gap, we propose KG-EGV, a versatile framework leveraging LLMs to perform KG-based tasks. KG-EGV consists of four core steps: sentence segmentation, graph retrieval, EGV, and backward updating, each designed to segment sentences, retrieve relevant KG components, and derive logical conclusions. EGV, a novel integrated framework for LLM inference, enables comprehensive reasoning beyond retrieval by synthesizing diverse evidence, which is often unattainable via retrieval alone due to noise or hallucinations. The framework incorporates six key stages: generation expansion, expansion evaluation, document re-ranking, re-ranking evaluation, answer generation, and answer verification. Within this framework, LLMs take on various roles, such as generator, re-ranker, evaluator, and verifier, collaboratively enhancing answer precision and logical coherence. By combining the strengths of retrieval-based and generation-based evidence, KG-EGV achieves greater flexibility and accuracy in evidence gathering and answer formulation. Extensive experiments on widely used benchmarks, including FactKG, MetaQA, NQ, WebQ, and TriviaQA, demonstrate that KG-EGV achieves state-of-the-art performance in answer accuracy and evidence quality, showcasing its potential to advance QA research and applications. Full article

In this paper, we describe our biomedical document retrieval system and answers extraction module, which is part of the biomedical question answering system. Approximately 26.5 million PubMed articles are indexed as a corpus with the Apache Lucene text search engine. Our proposed system consists of three parts. The first part is the question analysis module, which analyzes the question and enriches it with biomedical concepts related to its wording. The second part of the system is the document retrieval module. In this step, the proposed system is tested using different information retrieval models, like the Vector Space Model, Okapi BM25, and Query Likelihood. The third part is the document re-ranking module, which is responsible for re-arranging the documents retrieved in the previous step. For this study, we tested our proposed system with 6B training questions from the BioASQ challenge task. We obtained the best MAP score on the document retrieval phase when we used Query Likelihood with the Dirichlet Smoothing model. We used the sequential dependence model at the re-rank phase, but this model produced a worse MAP score than the previous phase. In similarity calculation, we included the Named Entity Recognition (NER), UMLS Concept Unique Identifiers (CUI), and UMLS Semantic Types of the words in the question to find the sentences containing the answer. Using this approach, we observed a performance enhancement of roughly 25% for the top 20 outcomes, surpassing another method employed in this study, which relies solely on textual similarity. Full article

Knowledge base question answering (KBQA) can be divided into two types according to the type of complexity: questions with constraints and questions with multiple hops of relationships. Previous work on knowledge base question answering have mostly focused on entities and relations. In a multihop question, it is insufficient to focus solely on topic entities and their relations since the relation between words also contains some important information. In addition, because the question contains constraints or multiple relationships, the information is difficult to capture, or the constraints are missed. In this paper, we applied a dependency structure to questions that capture relation information (e.g., constraint) between the words in question through a graph convolution network. The captured relation information is integrated into the question for re-encoding, and the information is used to generate and rank query graphs. Compared with existing sequence models and query graph generation models, our approach achieves a 0.8–3% improvement on two benchmark datasets. Full article

The 2019 European Open Data Directive identifies geospatial data as data that could have a major impact on human activities (high-value data, HVD) and advocates its provision as open data (OD), i.e., without barriers to access and re-use. Although Croatia has implemented OD policies to support the provision of open data, many geospatial data are still not available, or if available, their level of openness ranks Croatia lower than Slovenia and Serbia on some ranking lists. Benchmarking tools have proven to be a powerful tool in identifying barriers in OD. This paper, therefore, benchmarks the level of openness and provision of geospatial HVD in Croatia, Slovenia and Serbia, using the extended and modified Global Open Data Index methodology (GODI Plus). It is expected that this will provide an answer to the status of OD policies and government engagement in OD in Croatia and identify good OD practices among the three countries analyzed. Furthermore, the results will be a baseline benchmark for future HVD analyses. The results reveal low data openness for Croatia and Serbia, high data openness for Slovenia, and a low level of government engagement in all three proposed countries. Full article

Open-Domain Question Answering (Open-Domain QA) aims to answer any factoid questions from users. Recent progress in Open-Domain QA adopts the “retriever-reader” structure, which has proven effective. Retriever methods are mainly categorized as sparse retrievers and dense retrievers. In recent work, the dense retriever showed a stronger semantic interpretation than the sparse retriever. When training a dual-encoder dense retriever for document retrieval and reranking, there are two challenges: negative selection and a lack of training data. In this study, we make three major contributions to this topic: negative selection by query generation, data augmentation from negatives, and a passage evaluation method. We prove that the model performs better by focusing on false negatives and data augmentation in the Open-Domain QA passage rerank task. Our model outperforms other single dual-encoder rerankers over BERT-base and BM25 by 0.7 in MRR@10, achieving the highest Recall@50 and the max Recall@1000, which is restricted by the BM25 retrieval results. Full article

The paper concerns the role of the banking sector in renewable energy financing in Poland. The main goal of the paper is to provide recommendations for the banking sector in Poland, which can be used in the process of financing RES. The main methods used in the paper are the thorough analysis of the solutions used to finance RES in different countries and multivariate analysis of options presented on the ordinal scale. The first finding is the answer to the question of which financial instruments used by banks are the most effective in the financing of RES. It is based on the prepared ranking of different instruments used by banks in the process of renewable energy financing, by assessing the structure and value of required financing for renewable energy based on future scenarios. The second finding in the paper is the set of recommendations for the banking sector and policymakers as to financing renewable energy sources in Poland. The main conclusion is that renewable energy financing through the instruments available in the banking sector is efficient and is characterized by relatively low risk. Full article

Background: Alzheimer’s disease (AD) is the major cause of age-associated cognitive decline, and in the absence of effective therapeutics is progressive and ultimately fatal, creating a dire need for successful prevention and treatment strategies. We recently reported results of a successful proof-of-concept trial, using a personalized, precision medicine protocol, but whether such an approach is readily scalable is unknown. Objective: In the case of AD, there is not a single therapeutic that exerts anything beyond a marginal, unsustained, symptomatic effect. This suggests that the monotherapeutic approach of drug development for AD may not be an optimal one, at least when used alone. Using a novel, comprehensive, and personalized therapeutic system called ReCODE (reversal of cognitive decline), which proved successful in a small, proof-of-concept trial, we sought to determine whether the program could be scaled to improve cognitive and metabolic function in individuals diagnosed with subjective cognitive impairment, mild cognitive impairment, and early-stage AD. Methods: 255 individuals submitted blood samples, took the Montreal Cognitive Assessment (MoCA) test, and answered intake questions. Individuals who enrolled in the ReCODE program had consultations with clinical practitioners, and explanations of the program were provided. Participants had follow-up visits that included education regarding diet, lifestyle choices, medications, supplements, repeat blood sample analysis, and MoCA testing between 2 and 12 months after participating in the ReCODE program. Pre- and post-treatment measures were compared using the non-parametric Wilcoxon signed rank test. Results and Conclusions: By comparing baseline to follow-up testing, we observed that MoCA scores either significantly improved or stabilized in the entire participant pool—results that were not as successful as those in the proof-of-concept trial, but more successful than anti-amyloid therapies—and other risk factors including blood glucose, high-sensitivity C-reactive protein, HOMA-IR, and vitamin D significantly improved in the participant pool. Our findings provide evidence that a multi-factorial, comprehensive, and personalized therapeutic program designed to mitigate AD risk factors can improve risk factor scores and stabilize or reverse the decline in cognitive function. Since superior results were obtained in the proof-of-concept trial, which was conducted by a small group of highly trained and experienced physicians, it is possible that results from the use of this personalized approach would be enhanced by further training and experience of the practicing physicians. Nonetheless, the current results provide further support indicating the potential of such an approach for the prevention and reversal of cognitive decline. Full article

Machine Reading Comprehension (MRC) research concerns how to endow machines with the ability to understand given passages and answer questions, which is a challenging problem in the field of natural language processing. To solve the Chinese MRC task efficiently, this paper proposes an Improved Extraction-based Reading Comprehension method with Answer Re-ranking (IERC-AR), consisting of a candidate answer extraction module and a re-ranking module. The candidate answer extraction module uses an improved pre-training language model, RoBERTa-WWM, to generate precise word representations, which can solve the problem of polysemy and is good for capturing Chinese word-level features. The re-ranking module re-evaluates candidate answers based on a self-attention mechanism, which can improve the accuracy of predicting answers. Traditional machine-reading methods generally integrate different modules into a pipeline system, which leads to re-encoding problems and inconsistent data distribution between the training and testing phases; therefore, this paper proposes an end-to-end model architecture for IERC-AR to reasonably integrate the candidate answer extraction and re-ranking modules. The experimental results on the Les MMRC dataset show that IERC-AR outperforms state-of-the-art MRC approaches. Full article

Recently, the performance of machine-reading and comprehension (MRC) systems has been significantly enhanced. However, MRC systems require high-performance text retrieval models because text passages containing answer phrases should be prepared in advance. To improve the performance of text retrieval models underlying MRC systems, we propose a re-ranking model, based on artificial neural networks, that is composed of a query encoder, a passage encoder, a phrase modeling layer, an attention layer, and a similarity network. The proposed model learns degrees of associations between queries and text passages through dot products between phrases that constitute questions and passages. In experiments with the MS-MARCO dataset, the proposed model demonstrated higher mean reciprocal ranks (MRRs), 0.8%p–13.2%p, than most of the previous models, except for the models based on BERT (a pre-trained language model). Although the proposed model demonstrated lower MRRs than the BERT-based models, it was approximately 8 times lighter and 3.7 times faster than the BERT-based models. Full article

Recent advances in deep neural networks, language modeling and language generation have introduced new ideas to the field of conversational agents. As a result, deep neural models such as sequence-to-sequence, memory networks, and the Transformer have become key ingredients of state-of-the-art dialog systems. While those models are able to generate meaningful responses even in unseen situations, they need a lot of training data to build a reliable model. Thus, most real-world systems have used traditional approaches based on information retrieval (IR) and even hand-crafted rules, due to their robustness and effectiveness, especially for narrow-focused conversations. Here, we present a method that adapts a deep neural architecture from the domain of machine reading comprehension to re-rank the suggested answers from different models using the question as a context. We train our model using negative sampling based on question–answer pairs from the Twitter Customer Support Dataset. The experimental results show that our re-ranking framework can improve the performance in terms of word overlap and semantics both for individual models as well as for model combinations. Full article

Web Search Engines provide a huge number of answers in response to a user query, many of which are not relevant, whereas some of the most relevant ones may not be found. In the literature several approaches have been proposed in order to help a user to find the information relevant to his/her real needs on the Web. To achieve this goal the individual Information Behavior can been analyzed to ’keep’ track of the user’s interests. Keeping information is a type of Information Behavior, and in several works researchers have referred to it as the study on what people do during a search on the Web. Generally, the user’s actions (e.g., how the user moves from one Web page to another, or her/his download of a document, etc.) are recorded in Web logs. This paper reports on research activities which aim to exploit the information extracted from Web logs (or query logs) in personalized user ontologies, with the objective to support the user in the process of discovering Web information relevant to her/his information needs. Personalized ontologies are used to improve the quality of Web search by applying two main techniques: query reformulation and re-ranking of query evaluation results. In this paper we analyze various methodologies presented in the literature aimed at using personalized ontologies, defined on the basis of the observation of Information Behaviour to help the user in finding relevant information. Full article