Search Results (105)

Document-level Event Argument Extraction (DEAE) aims to extract structural event knowledge composed of arguments and roles beyond the sentence level. Existing methods mainly focus on designing prompts and using Abstract Meaning Representation (AMR) graph structure as additional features to enrich event argument representation. However, two challenges still remain: (1) the long-range dependency between event trigger and event arguments and (2) the distracting context in the document towards an event that can mislead the argument classification. To address these issues, we propose a novel document-level event argument extraction model named AMR Parser and Sparse Representation (APSR). Specifically, APSR sets inter- and intra-sentential encoders to capture the contextual information in different scopes. Especially, in the intra-sentential encoder, APSR designs three types of sparse event argument attention mechanisms to extract the long-range dependency. Then, APSR constructs AMR semantic graphs, which capture the interactions among concepts well. Finally, APSR fuses the inter- and intra-sentential representations and predicts what role a candidate span plays. Experimental results on the RAMS and WikiEvents datasets demonstrate that APSR achieves a superior performance compared with competitive baselines in terms of F1 by $1.27\%$ and $3.12\%$, respectively. Full article

This article contributes the first corpus-based study of possession in Māori, the indigenous language of Aotearoa New Zealand. Like most Polynesian languages, Māori has a dual possessive system involving a choice between the so-called A and O categories. While Māori grammars describe these categories in terms of the inherent semantic relationship between the possessum and possessor, there have been no large-scale corpus analyses demonstrating their use in natural contexts. Social media provide invaluable opportunities for such linguistic studies, capturing contemporary language use while alleviating the burden of gathering data through traditional means. We operationalise semantic distinctions to investigate possession in Māori-language tweets, focusing on the [possessum a/o possessor] construction (e.g., te tīmatanga o te wiki ‘the beginning of the week’). In our corpus comprising 2500 tweets produced by more than 200 individuals, we find that users leverage a wide array of noun types encompassing many different semantic relationships. We observe not only the expected predominance of the O category, but also a tendency for examples described by Māori grammars as A-marked to instead be O-marked (59%). Although the A category persists in the corpus, our findings suggest that language change could be underway. Our primary dataset can be explored interactively online. Full article

Drug induced fatty liver disease (DIFLD) is a form of drug-induced liver injury (DILI), which can also be included in the more general metabolic dysfunction-associated steatotic liver disease (MASLD), which specifically refers to the accumulation of fat in the liver unrelated to alcohol intake. A bi-directional relationship between DILI and MASLD is likely to exist: while certain drugs can cause MASLD by acting as pro-steatogenic factors, MASLD may make hepatocytes more vulnerable to drugs. Having a pre-existing MASLD significantly heightens the likelihood of experiencing DILI from certain medications. Thus, the prevalence of steatosis within DILI may be biased by pre-existing MASLD, and it can be concluded that the genuine true incidence of DIFLD in the general population remains unknown. In certain individuals, drug-induced steatosis is often accompanied by concomitant injury mechanisms such as oxidative stress, cell death, and inflammation, which leads to the development of drug-induced steatohepatitis (DISH). DISH is much more severe from the clinical point of view, has worse prognosis and outcome, and resembles MASH (metabolic-associated steatohepatitis), as it is associated with inflammation and sometimes with fibrosis. A literature review of clinical case reports allowed us to examine and evaluate the clinical features of DIFLD and their association with specific drugs, enabling us to propose a classification of DIFLD drugs based on clinical outcomes and pathological severity: Group 1, drugs with low intrinsic toxicity (e.g., ibuprofen, naproxen, acetaminophen, irinotecan, methotrexate, and tamoxifen), but expected to promote/aggravate steatosis in patients with pre-existing MASLD; Group 2, drugs associated with steatosis and only occasionally with steatohepatitis (e.g., amiodarone, valproic acid, and tetracycline); and Group 3, drugs with a great tendency to transit to steatohepatitis and further to fibrosis. Different mechanisms may be in play when identifying drug mode of action: (1) inhibition of mitochondrial fatty acid β-oxidation; (2) inhibition of fatty acid transport across mitochondrial membranes; (3) increased de novo lipid synthesis; (4) reduction in lipid export by the inhibition of microsomal triglyceride transfer protein; (5) induction of mitochondrial permeability transition pore opening; (6) dissipation of the mitochondrial transmembrane potential; (7) impairment of the mitochondrial respiratory chain/oxidative phosphorylation; (8) mitochondrial DNA damage, degradation and depletion; and (9) nuclear receptors (NRs)/transcriptomic alterations. Currently, the majority of, if not all, adverse outcome pathways (AOPs) for steatosis in AOP-Wiki highlight the interaction with NRs or transcription factors as the key molecular initiating event (MIE). This perspective suggests that chemical-induced steatosis typically results from the interplay between a chemical and a NR or transcription factors, implying that this interaction represents the primary and pivotal MIE. However, upon conducting this exhaustive literature review, it became evident that the current AOPs tend to overly emphasize this interaction as the sole MIE. Some studies indeed support the involvement of NRs in steatosis, but others demonstrate that such NR interactions alone do not necessarily lead to steatosis. This view, ignoring other mitochondrial-related injury mechanisms, falls short in encapsulating the intricate biological mechanisms involved in chemically induced liver steatosis, necessitating their consideration as part of the AOP’s map road as well. Full article

The Tibetan Plateau (TP) region contains maximal alpine grassland ecology at the mid-latitudes. This region is also recognized as an ecologically fragile and sensitive area under the effects of global warming. Regional climate modeling and ecosystem research depend on accurate land cover (LC) information. In order to obtain accurate LC information over the TP, the reliability and precision of five moderate/high-resolution LC products (MCD12Q1, C3S-LC, GlobeLand30, GLC_FCS30, and ESA2020 in 2020) were analyzed and evaluated in this study. The different LC products were compared with each other in terms of areal/spatial consistency and assessed with four reference sample datasets (Geo-Wiki, GLCVSS, GOFC-GOLD, and USGS) using the confusion matrix method for accuracy evaluation over the TP. Based on the paired comparison of these five LC datasets, all five LC products show that grass is the major land cover type on the TP, but the range of grass coverage identified by the different products varies noticeably, from 43.35% to 65.49%. The fully consistent spatial regions account for 43.72% of the entire region of the TP, while, in the transition area between grass and bare soil, there is still a large area of medium-to-low consistency. In addition, a comparison of LC datasets using integrated reference datasets shows that the overall accuracies of MCD12Q1, C3S-LC, GlobeLand30, GLC_FCS30, and ESA2020 are 54.29%, 49.32%, 53.03%, 53.73%, and 60.11%, respectively. The producer accuracy of the five products is highest for grass, while glaciers have the most reliable and accurate characteristics among all LC products for users. These findings provide valuable insights for the selection of rational and appropriate LC datasets for studying land-atmosphere interactions and promoting ecological preservation in the TP. Full article

In a multi-document summarization task, if the user can decide on the summary topic, the generated summary can better align with the reader’s specific needs and preferences. This paper addresses the issue of overly general content generation by common multi-document summarization models and proposes a topic-oriented multi-document summarization (TOMDS) approach. The method is divided into two stages: extraction and abstraction. During the extractive stage, it primarily identifies and retrieves paragraphs relevant to the designated topic, subsequently sorting them based on their relevance to the topic and forming an initial subset of documents. In the abstractive stage, building upon the transformer architecture, the process includes two parts: encoding and decoding. In the encoding part, we integrated an external discourse parsing module that focuses on both micro-level within-paragraph semantic relationships and macro-level inter-paragraph connections, effectively combining these with the implicit relationships in the source document to produce more enriched semantic features. In the decoding part, we incorporated a topic-aware attention mechanism that dynamically zeroes in on information pertinent to the chosen topic, thus guiding the summary generation process more effectively. The proposed model was primarily evaluated using the standard text summary dataset WikiSum. The experimental results show that our model significantly enhanced the thematic relevance and flexibility of the summaries and improved the accuracy of grammatical and semantic comprehension in the generated summaries. Full article

Emotional propagation research aims to discover and show the laws of opinion evolution in social networks. The short-term observation of the emotional propagation process for a predetermined time window ignores situations in which users with different emotions compete over a long diffusion time. To that end, we propose a dynamic emotional propagation model based on an independent cascade. The proposed model is inspired by the interpretable factors of the reinforced Poisson process, portraying the “rich-get-richer” phenomenon within a social network. Specifically, we introduce a time-decay mechanism to illustrate the change in influence over time. Meanwhile, we propose an emotion-exciting mechanism allowing prior users to affect the emotions of subsequent users. Finally, we conduct experiments on an artificial network and two real-world datasets—Wiki, with 7194 nodes, and Bitcoin-OTC, with 5881 nodes—to verify the effectiveness of our proposed model. The proposed method improved the F1-score by 3.5% and decreased the MAPE by 0.059 on the Wiki dataset. And the F1-score improved by 0.4% and the MAPE decreased by 0.013 on the Bitcoin-OTC dataset. In addition, the experimental results indicate a phenomenon of emotions in social networks tending to converge under the influence of opinion leaders after a long enough time. Full article

The COVID-19 disease, caused by the Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2), emerged in late 2019 and rapidly spread worldwide, becoming a pandemic that infected millions of people and caused significant deaths. COVID-19 continues to be a major threat, and there is a need to deepen our understanding of the virus and its mechanisms of infection. To study the cellular responses to SARS-CoV-2 infection, we performed an RNA sequencing of infected vs. uninfected Calu-3 cells. Total RNA was extracted from infected (0.5 MOI) and control Calu-3 cells and converted to cDNA. Sequencing was performed, and the obtained reads were quality-analyzed and pre-processed. Differential expression was assessed with the EdgeR package, and functional enrichment was performed in EnrichR for Gene Ontology, KEGG pathways, and WikiPathways. A total of 1040 differentially expressed genes were found in infected vs. uninfected Calu-3 cells, of which 695 were up-regulated and 345 were down-regulated. Functional enrichment analyses revealed the predominant up-regulation of genes related to innate immune response, response to virus, inflammation, cell proliferation, and apoptosis. These transcriptional changes following SARS-CoV-2 infection may reflect a cellular response to the infection and help to elucidate COVID-19 pathogenesis, in addition to revealing potential biomarkers and drug targets. Full article

Collaborative learning has been widely acknowledged as a successful teaching method within the education field, with research indicating its positive impact on student outcomes. During the COVID-19 pandemic, when all courses transitioned online due to lockdown measures, many universities employed learning management systems to facilitate continued group work among students. However, forming effective student groups remained challenging, particularly given the large number of enrolled students. To address this issue, this study proposes the application of an artificial intelligence (machine learning) solution to automatically group students based on their behaviours and interactions within an e-learning environment. This paper explores the potential of machine learning (ML) algorithms in assisting educators to create heterogeneous groups, considering various student attributes, such as behaviour and performance, to optimise collaborative learning outcomes. Students’ performance within a module was compared using a wiki activity that employed group work over the course of two academic years. In the first experiment, groups were formed randomly, while in the second experiment, students with similar behaviours were firstly identified using a clustering algorithm and then organised by an additional algorithm into heterogeneous groups. The results demonstrate the efficacy of the machine learning solution compared to the random approach in assisting educators with group formation for a collaborative activity such as the wiki, confirmed by a comparative analysis showing an improvement in student performance and satisfaction. This research contributes to the advancement of online education through the creation of more effective group dynamics using machine learning algorithms, thereby improving overall student learning. Full article

This study’s main goal is to create a useful software application for finding and classifying fine art photos in museums and art galleries. There is an increasing need for tools to swiftly analyze and arrange art collections based on their artistic styles as a result of the digitization of art collections. To increase the accuracy of the style categorization, the suggested technique involves two parts. The input image is split into five sub-patches in the first stage. A DCNN that has been particularly trained for this task is then used to classify each patch individually. A decision-making module using a shallow neural network is part of the second phase. Probability vectors acquired from the first-phase classifier are used to train this network. The results from each of the five patches are combined in this phase to deduce the final style classification for the input image. One key advantage of this approach is employing probability vectors rather than images, and the second phase is trained separately from the first. This helps compensate for any potential errors made during the first phase, improving accuracy in the final classification. To evaluate the proposed method, six various already-trained CNN models, namely AlexNet, VGG-16, VGG-19, GoogLeNet, ResNet-50, and InceptionV3, were employed as the first-phase classifiers. The second-phase classifier was implemented as a shallow neural network. By using four representative art datasets, experimental trials were conducted using the Australian Native Art dataset, the WikiArt dataset, ILSVRC, and Pandora 18k. The findings showed that the recommended strategy greatly surpassed existing methods in terms of style categorization accuracy and precision. Overall, the study assists in creating efficient software systems for analyzing and categorizing fine art images, making them more accessible to the general public through digital platforms. Using pre-trained models, we were able to attain an accuracy of 90.7. Our model performed better with a higher accuracy of 96.5 as a result of fine-tuning and transfer learning. Full article

Semantic web technologies (SWT) represent data in a format that is easier for machines to understand. Validating the knowledge in data graphs created using SWT is critical to ensure that the axioms accurately represent the so-called “real” world. However, data graph validation is a significant challenge in the semantic web domain. The Shapes Constraint Language (SHACL) is the latest W3C standard developed with the goal of validating data-graphs. SHACL (pronounced as shackle) is a relatively new standard and hitherto has predominantly been employed to validate generic data graphs like WikiData and DBPedia. In generic data graphs, the name of a class does not affect the shape of a class, but this is not the case with biomedical ontology data graphs. The shapes of classes in biomedical ontology data graphs are highly influenced by the names of the classes, and the SHACL shape creation methods developed for generic data graphs fail to consider this characteristic difference. Thus, the existing SHACL shape creation methods do not perform well for domain-specific biomedical ontology data graphs. Maintaining the quality of biomedical ontology data graphs is crucial to ensure accurate analysis in safety-critical applications like Electronic Health Record (EHR) systems referencing such data graphs. Thus, in this work, we present a novel method to create enhanced SHACL shapes that consider the aforementioned characteristic difference to better validate biomedical ontology data graphs. We leverage the knowledge available from lexical auditing techniques for biomedical ontologies and incorporate this knowledge to create smart SHACL shapes. We also create SHACL shapes (baseline SHACL graph) without incorporating the lexical knowledge of the class names, as is performed by existing methods, and compare the performance of our enhanced SHACL shapes with the baseline SHACL shapes. The results demonstrate that the enhanced SHACL shapes augmented with lexical knowledge of the class names identified 176 violations which the baseline SHACL shapes, void of this lexical knowledge, failed to detect. Thus, the enhanced SHACL shapes presented in this work significantly improve the validation performance of biomedical ontology data graphs, thereby reducing the errors present in such data graphs and ensuring safe use in the life-critical applications referencing them. Full article

Background: Osteoporosis, which is a bone disease, is characterized by low bone mineral density and an increased risk of fractures. The heel bone mineral density is often used as a representative measure of overall bone mineral density. Lipid metabolism, which includes processes such as fatty acid metabolism, glycerol metabolism, inositol metabolism, bile acid metabolism, carnitine metabolism, ketone body metabolism, sterol and steroid metabolism, etc., may have an impact on changes in bone mineral density. While some studies have reported correlations between lipid metabolism and heel bone mineral density, the overall causal relationship between metabolites and heel bone mineral density remains unclear. Objective: to investigate the causal relationship between lipid metabolites and heel bone mineral density using two-sample Mendelian randomization analysis. Methods: Summary-level data from large-scale genome-wide association studies were extracted to identify genetic variants linked to lipid metabolite levels. These genetic variants were subsequently employed as instrumental variables in Mendelian randomization analysis to estimate the causal effects of each lipid metabolite on heel bone mineral density. Furthermore, metabolites that could potentially be influenced by causal relationships with bone mineral density were extracted from the KEGG and WikiPathways databases. The causal associations between these downstream metabolites and heel bone mineral density were then examined. Lastly, a sensitivity analysis was conducted to evaluate the robustness of the results and address potential sources of bias. Results: A total of 130 lipid metabolites were analyzed, and it was found that acetylcarnitine, propionylcarnitine, hexadecanedioate, tetradecanedioate, myo-inositol, 1-arachidonoylglycerophosphorine, 1-linoleoylglycerophoethanolamine, and epiandrosterone sulfate had a causal relationship with heel bone mineral density (p < 0.05). Furthermore, our findings also indicate an absence of causal association between the downstream metabolites associated with the aforementioned metabolites identified in the KEGG and WikiPathways databases and heel bone mineral density. Conclusion: This work supports the hypothesis that lipid metabolites have an impact on bone health through demonstrating a causal relationship between specific lipid metabolites and heel bone mineral density. This study has significant implications for the development of new strategies to osteoporosis prevention and treatment. Full article

With the development of deep learning, image synthesis has achieved unprecedented achievements in the past few years. Image synthesis models, represented by diffusion models, demonstrated stable and high-fidelity image generation. However, the traditional diffusion model computes in pixel space, which is memory-heavy and computing-heavy. Therefore, to ease the expensive computing and improve the accessibility of diffusion models, we train the diffusion model in latent space. In this paper, we are devoted to creating novel paintings from existing paintings based on powerful diffusion models. Because the cross-attention layer is adopted in the latent diffusion model, we can create novel paintings with conditional text prompts. However, direct training of the diffusion model on the limited dataset is non-trivial. Therefore, inspired by the transfer learning, we train the diffusion model with the pre-trained weights, which eases the training process and enhances the image synthesis results. Additionally, we introduce the GPT-2 model to expand text prompts for detailed image generation. To validate the performance of our model, we train the model on paintings of the specific artist from the dataset WikiArt. To make up for missing image context descriptions of the WikiArt dataset, we adopt a pre-trained language model to generate corresponding context descriptions automatically and clean wrong descriptions manually, and we will make it available to the public. Experimental results demonstrate the capacity and effectiveness of the model. Full article

Pre-trained language model-based neural rankers have been widely applied in information retrieval (IR). However, the robustness issue of current IR models has not received sufficient attention, which could significantly impact the user experience in practical applications. In this study, we focus on the defensive ability of IR models against query attacks while guaranteeing their retrieval performance. We discover that improving the robustness of IR models not only requires a focus on model architecture and training methods but is also closely related to the quality of data. Different from previous research, we use large language models (LLMs) to generate query variations with the same intent, which exhibit richer and more realistic expressions while maintaining consistent query intent. Based on LLM-generated query variations, we propose a novel contrastive training framework that substantially enhances the robustness of IR models to query perturbations. Specifically, we combine the contrastive loss in the representation space of query variations with the ranking loss in the retrieval training stage to improve the model’s ability to understand the underlying semantic information of queries. Experimental results on two public datasets, WikiQA and ANTIQUE, demonstrate that the proposed contrastive training approach effectively improves the robustness of models facing query attack scenarios while outperforming baselines in retrieval performance. Compared with the best baseline approach, the improvements in average robustness performance of Reranker IR models are 24.9%, 26.5%, 27.0%, and 75.0% on WikiQA and 8.7%, 1.9%, 6.3%, and 13.6% on ANTIQUE, in terms of the MAP (Mean Average Precision), MRR (Mean Reciprocal Rank), nDCG@10 (Normalized Discounted Cumulative Gain) and P@10 (Precision), respectively. Full article

Considering the complexity of entity pair relations and the information contained in the target neighborhood in few-shot knowledge graphs (KG), existing few-shot KG completion methods generally suffer from insufficient relation representation learning capabilities and neglecting the contextual semantics of entities. To tackle the above problems, we propose a Few-shot Relation Learning-based Knowledge Graph Completion model (FRL-KGC). First, a gating mechanism is introduced during the aggregation of higher-order neighborhoods of entities in formation, enriching the central entity representation while reducing the adverse effects of noisy neighbors. Second, during the relation representation learning stage, a more accurate relation representation is learned by using the correlation between entity pairs in the reference set. Finally, an LSTM structure is incorporated into the Transformer learner to enhance its ability to learn the contextual semantics of entities and relations and predict new factual knowledge. We conducted comparative experiments on the publicly available NELL-One and Wiki-One datasets, comparing FRL-KGC with six few-shot knowledge graph completion models and five traditional knowledge graph completion models for five-shot link prediction. The results showed that FRL-KGC outperformed all comparison models in terms of MRR, Hits@10, Hits@5, and Hits@1 metrics. Full article

The recent advancements in big data and natural language processing (NLP) have necessitated proficient text mining (TM) schemes that can interpret and analyze voluminous textual data. Text summarization (TS) acts as an essential pillar within recommendation engines. Despite the prevalent use of abstractive techniques in TS, an anticipated shift towards a graph-based extractive TS (ETS) scheme is becoming apparent. The models, although simpler and less resource-intensive, are key in assessing reviews and feedback on products or services. Nonetheless, current methodologies have not fully resolved concerns surrounding complexity, adaptability, and computational demands. Thus, we propose our scheme, GETS, utilizing a graph-based model to forge connections among words and sentences through statistical procedures. The structure encompasses a post-processing stage that includes graph-based sentence clustering. Employing the Apache Spark framework, the scheme is designed for parallel execution, making it adaptable to real-world applications. For evaluation, we selected 500 documents from the WikiHow and Opinosis datasets, categorized them into five classes, and applied the recall-oriented understudying gisting evaluation (ROUGE) parameters for comparison with measures ROUGE-1, 2, and L. The results include recall scores of 0.3942, 0.0952, and 0.3436 for ROUGE-1, 2, and L, respectively (when using the clustered approach). Through a juxtaposition with existing models such as BERTEXT (with 3-gram, 4-gram) and MATCHSUM, our scheme has demonstrated notable improvements, substantiating its applicability and effectiveness in real-world scenarios. Full article