Search Results (42)

With the accelerated digital transformation in education, the efficient integration of massive multimodal instructional resources and the support for interactive question answering (QA) remains a prominent challenge. This study introduces Multimodal Disciplinary Knowledge-Augmented Generation (MDKAG), a framework integrating retrieval-augmented generation (RAG) with a multimodal disciplinary knowledge graph (MDKG). MDKAG first extracts high-precision entities from digital textbooks, lecture slides, and classroom videos by using the Enhanced Representation through Knowledge Integration 3.0 (ERNIE 3.0) model and then links them into a graph that supports fine-grained retrieval. At inference time, the framework retrieves graph-adjacent passages, integrates multimodal data, and feeds them into a large language model (LLM) to generate context-aligned answers. An answer-verification module checks semantic overlap and entity coverage to filter hallucinations and triggers incremental graph updates when new concepts appear. Experiments on three university courses show that MDKAG reduces hallucination rates by up to 23% and increases answer accuracy by 11% over text-only RAG and knowledge-augmented generation (KAG) baselines, demonstrating strong adaptability across subject domains. The results indicate that MDKAG offers an effective route for scalable knowledge organization and reliable interactive QA in education. Full article

The present paper reports for the first time the comparison of radar-derived eddy dissipation rate (EDR) and vertical velocity with measurements from six aircraft for an intense squall line crossing Hong Kong. The study objectives are three-fold: (i) to characterise the structural dynamics of the intense squall line; (ii) to identify the dynamical change in EDR and vertical velocity during its eastward propagation across Hong Kong with a view to gaining insight into the intensity change of the squall line and the severity of its impact on aircraft flying near it; (iii) to carry out quantitative comparison of EDR and vertical velocity derived from remote sensing instruments, i.e., weather radars and in situ measurements from aircraft, so that the quality of the former dataset can be evaluated by the latter. During the passage of the squall line and taking reference of the radar reflectivity, vertical circulation and the subsiding flow at the rear, it appeared to be weakening in crossing over Hong Kong, possibly due to land friction by terrain and urban morphology. This is also consistent with the maximum gusts recorded by the dense network of ground-based anemometers in Hong Kong. However, from the EDR and the vertical velocity of the aircraft, the weakening trend was not very apparent, and rather severe turbulence was still recorded by the aircraft flying through the squall line into the region with stratiform precipitation when the latter reached the eastern coast of Hong Kong. In general, the radar-based and the aircraft-based EDRs are consistent with each other. The radar-retrieved maximum vertical velocity may be smaller in magnitude at times, possibly arising from the limited spatial and temporal resolutions of the aircraft data. The results of this paper could be a useful reference for the development of radar-based turbulence products for aviation applications. Full article

Fusion-in-Decoder (FiD), a prominent retrieval-augmented generation model, has demonstrated outstanding performance in open-domain question answering by effectively leveraging multiple passages. However, processing multiple passages significantly increases computational costs at both encoder and decoder components. In particular, in Long-Form Question Answering (LFQA) scenarios, the decoder’s cross-attention computation scales proportionally with the length of the generated answer, severely impacting the overall inference speed. In this paper, we propose a novel dynamic token pruning mechanism to alleviate the computational bottleneck of the FiD decoder. Our method selectively identifies and removes tokens predicted to have low contributions to answer generation by jointly considering their contextual information and attention scores within the FiD encoder. The resulting pruned representations are then passed to the decoder, significantly reducing the cross-attention computations and thereby accelerating the overall inference process. Experimental evaluations on two LFQA benchmarks, ASQA and CLAPNQ, demonstrate that the proposed method achieves up to a 1.74-fold speed-up while maintaining minimal degradation in answer quality, effectively enhancing computational efficiency compared to the original FiD model. Full article

Evaluating query-passage relevance is a crucial task in information retrieval (IR), where the performance of large language models (LLMs) greatly depends on the quality of prompts. Current prompt optimization methods typically require multiple candidate generations or iterative refinements, resulting in significant computational overhead and limited practical applicability. In this paper, we propose a novel prompt optimization method that leverages LLM-based confusion matrix feedback to efficiently optimize prompts for the relevance evaluation task. Unlike previous approaches, our method systematically analyzes LLM predictions—both correct and incorrect—using a confusion matrix, enabling prompt refinement through a single-step update. Our experiments in realistic IR scenarios demonstrate that our method achieves competitive or superior performance compared to existing methods while drastically reducing computational costs, highlighting its potential as a practical and scalable solution. Full article

In the field of Information Retrieval (IR), user-submitted keyword queries often fail to accurately represent users’ true search intent. With the rapid advancement of artificial intelligence, particularly in natural language processing (NLP), query expansion (QE) based on large language models (LLMs) has emerged as a key strategy for improving retrieval effectiveness. However, such methods often introduce query topic drift, which negatively impacts retrieval accuracy and efficiency. To address this issue, this study proposes an LLM-based QE framework that incorporates a Gaussian kernel-enhanced semantic space for dense retrieval. Specifically, the model first employs LLMs to expand the semantic dimensions of the initial query, generating multiple query representations. Then, by introducing a Gaussian kernel semantic space, it captures deep semantic relationships among these query vectors, refining their semantic distribution to better represent the original query’s intent. Finally, the ColBERTv2 model is utilized to retrieve documents based on the enhanced query representations, enabling precise relevance assessment and improving retrieval performance. To validate the effectiveness of the proposed approach, extensive empirical evaluations were conducted on the MS MARCO passage ranking dataset. The model was systematically assessed using key metrics, including MAP, NDCG@10, MRR@10, and Recall@1000. Experimental results demonstrate that the proposed method outperforms existing approaches across multiple metrics, significantly improving retrieval precision while effectively mitigating query drift, offering a novel approach for building efficient QE mechanisms. Full article

In this paper various LLMs are tested in a specific domain using a Retrieval-Augmented Generation (RAG) system. The study focuses on the performance and behavior of the models and was conducted in Spanish. A questionnaire based on The Bible, which consists of questions that vary in complexity of reasoning, was created in order to evaluate the reasoning capabilities of each model. The RAG system matches a question with the most similar passage from The Bible and feeds the pair to each LLM. The evaluation aims to determine whether each model can reason solely with the provided information or if it disregards the instructions given and makes use of its pretrained knowledge. Full article

The search for environmental data typically involves lexical approaches, where query terms are matched with metadata records based on measures of term frequency. In contrast, dense retrieval approaches employ language models to comprehend the context and meaning of a query and provide relevant search results. However, for environmental data, this has not been researched and there are no corpora or evaluation datasets to fine-tune the models. This study demonstrates the adaptation of dense retrievers to the domain of climate-related scientific geodata. Four corpora containing text passages from various sources were used to train different dense retrievers. The domain-adapted dense retrievers are integrated into the search architecture of a standard metadata catalogue. To improve the search results further, we propose a spatial re-ranking stage after the initial retrieval phase to refine the results. The evaluation demonstrates superior performance compared to the baseline model commonly used in metadata catalogues (BM25). No clear trends in performance were discovered when comparing the results of the dense retrievers. Therefore, further investigation aspects are identified to finally enable a recommendation of the most suitable corpus composition. Full article

This study evaluates the capability of an unconstrained tomographic algorithm to capture 3D water vapor density variability throughout 2017 in Houston, U.S. The algorithm relies solely on Global Navigation Satellite System (GNSS) observations and does not require an initial guess or other specific constraints regarding water vapor density variability within the tomographic domain. The test domain, featuring 9 km horizontal, 500 m vertical, and 30 min temporal resolutions, yielded remarkable results when compared to data retrieved from the ECMWF Reanalysis v5 (ERA5), regional Weather Research and Forecasting Model (WRF) data, and GNSS-Radio Occultation (RO). For the first time, a time series of Precipitable Water Vapor maps derived from the Interferometric Synthetic Aperture Radar (InSAR) technique was used to validate the spatially integrated water vapor computed by GNSS tomography. Tomographic results clearly indicate the passage of Hurricane Harvey, with integrated water vapor peaking at 60 kg/m² and increased humidity at altitudes up to 7.5 km. Our findings suggest that GNSS tomography holds promise as a reliable source of atmospheric water vapor data for various applications. Future enhancements may arise from denser and multi-constellation networks. Full article

In the field of Information Retrieval, biomedical question answering is a specialized task that focuses on answering questions related to medical and healthcare domains. The goal is to provide accurate and relevant answers to the posed queries related to medical conditions, treatments, procedures, medications, and other healthcare-related topics. Well-designed models should efficiently retrieve relevant passages. Early retrieval models can quickly retrieve passages but often with low precision. In contrast, recently developed Large Language Models can retrieve documents with high precision but at a slower pace. To tackle this issue, we propose a two-stage retrieval approach that initially utilizes BM25 for a preliminary search to identify potential candidate documents; subsequently, a Large Language Model is fine-tuned to evaluate the relevance of query–document pairs. Experimental results indicate that our approach achieves comparative performances on the BioASQ and the TREC-COVID datasets. Full article

Finding passages related to a sentence over a large collection of text documents is a fundamental task for claim verification and open-domain question answering. For instance, a common approach for verifying a claim is to extract short snippets of relevant text from a collection of reference documents and provide them as input to a natural language inference machine that determines whether the claim can be deduced or refuted. Available approaches struggle when several pieces of evidence from different documents need to be combined to make an inference, as individual documents often have a low relevance with the input and are therefore excluded. We propose GRAAL (GRAph-based retrievAL), a novel graph-based approach that outlines the relevant evidence as a subgraph of a large graph that summarizes the whole corpus. We assess the validity of this approach by building a large graph that represents co-occurring entity mentions on a corpus of Wikipedia pages and using this graph to identify candidate text relevant to a claim across multiple pages. Our experiments on a subset of FEVER, a popular benchmark, show that the proposed approach is effective in identifying short passages related to a claim from multiple documents. Full article

A typhoon disaster is a common meteorological disaster that seriously impacts natural ecology, social economy, and even human sustainable development. It is crucial to access the typhoon disaster information, and the corresponding disaster prevention and reduction strategies. However, traditional question and answering (Q&A) methods exhibit shortcomings like low information retrieval efficiency and poor interactivity. This makes it difficult to satisfy users’ demands for obtaining accurate information. Consequently, this work proposes a typhoon disaster knowledge Q&A approach based on LLM (T5). This method integrates two technical paradigms of domain fine-tuning and retrieval-augmented generation (RAG) to optimize user interaction experience and improve the precision of disaster information retrieval. The process specifically includes the following steps. First, this study selects information about typhoon disasters from open-source databases, such as Baidu Encyclopedia and Wikipedia. Utilizing techniques such as slicing and masked language modeling, we generate a training set and 2204 Q&A pairs specifically focused on typhoon disaster knowledge. Second, we continuously pretrain the T5 model using the training set. This process involves encoding typhoon knowledge as parameters in the neural network’s weights and fine-tuning the pretrained model with Q&A pairs to adapt the T5 model for downstream Q&A tasks. Third, when responding to user queries, we retrieve passages from external knowledge bases semantically similar to the queries to enhance the prompts. This action further improves the response quality of the fine-tuned model. Finally, we evaluate the constructed typhoon agent (Typhoon-T5) using different similarity-matching approaches. Furthermore, the method proposed in this work lays the foundation for the cross-integration of large language models with disaster information. It is expected to promote the further development of GeoAI. Full article

Gamma-aminobutyric acid (GABA), the primary inhibitory neurotransmitter in the central nervous system (CNS), plays a pivotal role in maintaining the delicate balance between inhibitory and excitatory neurotransmission. Dysregulation of the excitatory/inhibitory balance is implicated in various neurological and psychiatric disorders, emphasizing the critical role of GABA in disease-free brain function. The review examines the intricate interplay between the gut–brain axis and CNS function. The potential impact of dietary GABA on the brain, either by traversing the blood–brain barrier (BBB) or indirectly through the gut–brain axis, is explored. While traditional beliefs questioned GABA’s ability to cross the BBB, recent research challenges this notion, proposing specific transporter systems facilitating GABA passage. Animal studies provide some evidence that small amounts of GABA can cross the BBB but there is a lack of human data to support the role of transporter-mediated GABA entry into the brain. This review also explores GABA-containing food supplements, investigating their impact on brain activity and functions. The potential benefits of GABA supplementation on pain management and sleep quality are highlighted, supported by alterations in electroencephalography (EEG) brain responses following oral GABA intake. The comprehensive overview encompasses GABA’s sources in the diet, including brown rice, soy, adzuki beans, and fermented foods. GABA’s presence in various foods and supplements, its association with gut microbiota, and its potential as a therapeutic strategy for neurological disorders are thoroughly examined. The articles were retrieved through a systematic review of the databases: OVID, SCOPUS, and PubMed (keywords “GABA”, “oral GABA“, “sleep”, “cognition”, “neurodegenerative”, “blood-brain barrier”, “gut microbiota”, “supplements” and “therapeutic”, and by searching reference sections from identified studies and review articles). This review presents the relevant literature available on the topic and discusses the mechanisms, effects, and hypotheses that suggest oral GABA benefits range from neuroprotection to blood pressure control. The literature suggests that oral intake of GABA affects the brain illustrated by changes in EEG scans and cognitive performance, with evidence showing that GABA can have beneficial effects for multiple age groups and conditions. The potential clinical and research implications of utilizing GABA supplementation are vast, spanning a spectrum of diseases ranging from neurodegeneration to blood pressure regulation. Importantly, recommendations for the use of oral GABA should consider the dosage, formulation, and duration of treatment as well as potential side effects. Effects of GABA need to be more thoroughly investigated in robust clinical trials to validate efficacy to progress the development of alternative treatments for a variety of disorders. Full article

Analysis of the relationships among global stock markets is crucial for international investors, regulators, and policymakers, particularly during a crisis. Complex network theory was applied to analyze the relationship between global stock markets during the Russia–Ukraine war. Daily data from 55 stock markets from 6 August 2021 to 23 September 2023 were retrieved and used to investigate the changes in global stock market networks. The sample period was divided into 22 subsamples, using a 100-day rolling window rolled forward a trading month, and then long-range correlations based on distance matrices were calculated. These distance matrices were utilized to construct stock market networks. Moreover, minimum spanning trees (MSTs) were extracted from these financial networks for analytical purposes. Based on topological and structural analysis, we identified important/central nodes, distinct communities, vulnerable/stable nodes, and changes thereof with the escalation of war. The empirical findings reveal that the Russia–Ukraine war impacted the global stock markets’ network. However, its intensity varied with changes in the region and the passage of time due to the level of stock market integration and stage of war escalation, respectively. Stock markets of France, Germany, Canada, and Austria remained the most centrally connected within communities; surprisingly, the USA’s stock market is not on this list. Full article

The results of a field experiment devoted to observing slick-band shape variations occurring due to the action of heterogeneous currents related to the passage of internal waves are presented and analyzed on the basis of numerical simulation. The spatiotemporal structure of a train of five solitons of internal waves has been retrieved. Their evolution in the coastal area is demonstrated based on the analysis of propagation characteristics. It is shown that the first soliton, characterized by the higher values of amplitude and width, collapsed when entering shallow water near the observation platform. The parameters of an artificial slick band affected by the passage of internal waves are determined. It is shown that the direction and width of the slick band are related to the direction and magnitude of the upper-ocean horizontal current, which contains a component related to the internal wave. The results of numerical simulation are qualitatively and quantitatively consistent with experimental data at short distances from the platform. An analysis of the conditions responsible for different regimes of slick-band response to the upper-ocean currents generated by propagating internal waves has been performed. Full article

Sea ice motion (SIM) has significant implications for sea–air interactions, thermohaline circulation, and the development of the Arctic passage. This research proposes an improved SIM retrieval method from Fengyun-3D’s (FY-3D) microwave radiometer imager’s (MWRI) brightness temperature (T_b) data based on the modified classical maximum cross-correlation (MCC) method and the multisource data merging method. This study utilized buoy data to establish the search area range, applied distinct thresholds across various Arctic regions, and merged the buoy data, reanalysis wind data, and SIM retrieved from FY-3D/MWRI T_b data. In 2019, for the final Arctic SIM results retrieved from the MWRI 89 GHz and 36.5 GHz T_b data, the root-mean-square error (RMSE) and the mean average error (MAE) in the east–west direction were 2.07 cm/s and 1.38 cm/s and those in the north–south direction were 1.96 cm/s and 1.15 cm/s, compared to the ice-tethered profiler (ITP) data. Compared with the daily average data of the National Snow and Ice Data Center (NSIDC), the RMSE and MAE of the SIM results obtained in this study were 0.74 cm/s and 0.93 cm/s in the east–west direction, and 0.56 cm/s and 0.72 cm/s in the north–south direction, respectively. The monthly average of the SIM retrieved from the MWRI T_b data in this research also showed a good agreement with the monthly average of the NSIDC SIM product. The comparison showed that the MWRI T_b data could be used to retrieve the Arctic SIM, and the Arctic SIM retrieval method presented in this paper was accurate and general. Full article