Search Results (101)

Large-language-model (LLM) APIs demonstrate impressive reasoning capabilities, but their size, cost, and closed weights limit the deployment of knowledge-aware AI within biomedical research groups. At the other extreme, standard attention-based neural language models (SANLMs)—including encoder–decoder architectures such as Transformers, Gated Recurrent Units (GRUs), and Long Short-Term Memory (LSTM) networks—are computationally inexpensive. However, their capacity for semantic reasoning in noisy, open-vocabulary knowledge bases (KBs) remains unquantified. Therefore, we investigate whether compact SANLMs can (i) reason over hybrid OpenIE-derived KBs that integrate commonsense, general-purpose, and non-communicable-disease (NCD) literature; (ii) operate effectively on commodity GPUs; and (iii) exhibit semantic coherence as assessed through manual linguistic inspection. To this end, we constructed four training KBs by integrating ConceptNet (600k triples), a 39k-triple general-purpose OpenIE set, and an 18.6k-triple OpenNCDKB extracted from 1200 PubMed abstracts. Encoder–decoder GRU, LSTM, and Transformer models (1–2 blocks) were trained to predict the object phrase given the subject + predicate. Beyond token-level cross-entropy, we introduced the Meaning-based Selectional-Preference Test (MSPT): for each withheld triple, we masked the object, generated a candidate, and measured its surplus cosine similarity over a random baseline using word embeddings, with significance assessed via a one-sided t-test. Hyperparameter sensitivity (311 GRU/168 LSTM runs) was analyzed, and qualitative frame–role diagnostics completed the evaluation. Our results showed that all SANLMs learned effectively from the point of view of the cross entropy loss. In addition, our MSPT provided meaningful semantic insights: for the GRUs (256-dim, 2048-unit, 1-layer): mean similarity $\left({\mu }_{sts}\right)$ of 0.641 to the ground truth vs. 0.542 to the random baseline (gap 12.1%; $p<{10}^{-180}$). For the 1-block Transformer: ${\mu }_{sts}=0.551$ vs. $0.511$ (gap 4%; $p<{10}^{-25}$). While Transformers minimized loss and accuracy variance, GRUs captured finer selectional preferences. Both architectures trained within <24 GB GPU VRAM and produced linguistically acceptable, albeit over-generalized, biomedical assertions. Due to their observed performance, LSTM results were designated as baseline models for comparison. Therefore, properly tuned SANLMs can achieve statistically robust semantic reasoning over noisy, domain-specific KBs without reliance on massive LLMs. Their interpretability, minimal hardware footprint, and open weights promote equitable AI research, opening new avenues for automated NCD knowledge synthesis, surveillance, and decision support. Full article

Large-language models (LLMs) show promise for automating evidence synthesis, yet head-to-head evaluations remain scarce. We benchmarked five state-of-the-art LLMs—openai/o1-mini, x-ai/grok-2-1212, meta-llama/Llama-3.3-70B-Instruct, google/Gemini-Flash-1.5-8B, and deepseek/DeepSeek-R1-70B-Distill—on extracting protocol details from transcranial direct-current stimulation (tDCS) trials enrolling older adults. A multi-LLM ensemble pipeline ingested ClinicalTrials.gov records, applied a structured JSON schema, and generated comparable outputs from unstructured text. The pipeline retrieved 83 aging-related tDCS trials—roughly double the yield of a conventional keyword search. Across models, agreement was almost perfect for the binary field brain stimulation used (Fleiss κ ≈ 0.92) and substantial for the categorical primary target (κ ≈ 0.71). Numeric parameters such as stimulation intensity and session duration showed excellent consistency when explicitly reported (ICC 0.95–0.96); secondary targets and free-text duration phrases remained challenging (κ ≈ 0.61; ICC ≈ 0.35). An ensemble consensus (majority vote or averaging) resolved most disagreements and delivered near-perfect reliability on core stimulation attributes (κ = 0.94). These results demonstrate that multi-LLM ensembles can markedly expand trial coverage and reach expert-level accuracy on well-defined fields while still requiring human oversight for nuanced or sparsely reported details. The benchmark and open-source workflow set a solid baseline for future advances in prompt engineering, model specialization, and ensemble strategies aimed at fully automated evidence synthesis in neurostimulation research involving aging populations. Overall, the five-model multi-LLM ensemble doubled the number of eligible aging-related tDCS trials retrieved versus keyword searching and achieved near-perfect agreement on core stimulation parameters (κ ≈ 0.94), demonstrating expert-level extraction accuracy. Full article

“Martial arts” is a popular phrase in popular and academic discourse but notoriously difficult to define. This article addresses the challenge of defining martial arts, demonstrating the multifarious and sometimes contradictory nature of how the term is conceived in different contexts. Consulting a range of perspectives, the article is critical of essentialist positions in locating a permanent set of features common to all martial arts because definitions under such positions fail to consider these practices’ fluidity, hybridity, and historical evolution. Instead, the article advances a more pragmatic and contextual definition of martial arts, appealing to nominalism and diaeresis to build context-specific definitions appropriate for particular analytical or practical purposes. Acknowledging the diversity and complexity inherent in martial arts, the article suggests that scholars and practitioners can move beyond strict classification and engage in more fruitful discussions regarding these practices’ history, culture, and philosophy. Lastly, the article promotes a more inclusive and dynamic system that recognises both traditional and modern forms of martial arts without being constrained by the strictures of essentialist definitions. Full article

The theme of the current Special Issue, ‘Faith in Action: Examining the Power and Purpose of a Public Theology in Contemporary Society’, leaves a lot of scope for definition. Here, the theme is addressed via the lens of the public life and practice of a Christian congregation and its generative qualities pertaining to public trust. Such an approach contrasts with more formal public theologies which tend to favour intellectual or academic discourse. It will be argued that the life of a local faith community and its embodied public interface provides a better starting point as it can be seen as the most directly presupposed concrete and public context of the phrase ‘faith in action’. As an analytical conduit, the congregation is a neglected category for rendering a public theology, even as it lies at the heart of the Christian faith’s constitutive practice. Here, a particular form of public exchange drawn from local experience will be set against the background of the recent trend in ecclesiology to turn away from abstracted notions of the church towards more practice-orientated understandings. The dynamics of developing public trust will be considered through reflection on an oft repeated response offered from non-practicing observers of my local church’s work: ‘You do so much good in the community’. Although a seemingly simple (and positive) sentiment, such a comment in fact bears complex layers of meaning and subtextual inflections to which a minister and congregation might give heed in search of public trust. Full article

Sepsis is defined as a life-threatening organ dysfunction caused by a dysregulated host response to infection. The number of patients with sepsis requiring critical care admission is increasing. At the same time, overall mortality from sepsis is declining. With increasing survival to hospital discharge, there are an increasing number of sepsis survivors whose care needs shift from the acute to chronic care settings. Recently, the phrase “post-sepsis syndrome” has emerged to encompass the myriad of complications in patients recovering from sepsis. The aim of this narrative review is to provide a contemporary summary of the available literature on post-sepsis care and highlight areas of ongoing research. There are many incentives for improving the quality of survivorship following sepsis, including individual health-related outcomes (e.g., increased survival, enhanced physical and psychological health) and wider socio-economic benefits (e.g., reduced economic burden on the healthcare systems, reduced physical and psychological burden on carers, ability for individuals (and carers) to return to workforce). Modifiable factors influencing long-term outcomes can be in-hospital or after discharge, when primary care physicians play a pivotal role. Despite national and international guidance being available, this area has been under-recognised historically, despite its profoundly negative impact on both patients and their families or caregivers. Contributing factors likely include the lack of a formally recognised “disease” or pathology, the presence of challenging-to-treat symptoms such as fatigue, weakness and cognitive impairment, and the prevailing assumption that ongoing rehabilitation merely requires time. Our review will focus on the following areas: screening for new cognitive and physical impairments; optimisation of pre-existing comorbidities; transition to primary care; and palliative care. Primary care physicians may have a crucial role to play in improving outcomes in sepsis survivors, and candidate interventions include education on common complications of post-sepsis syndrome. Full article

The purpose of this bibliometric analysis is to understand what skill sets are needed for the entrepreneurial mindset, how the entrepreneurial mindset is practically operationalized, and where opportunities can be identified using the entrepreneurial mindset. The entrepreneurial mindset is crucial in entrepreneurship and innovation, leading to value creation, business development, and competitive advantage. The methodological approach involves a bibliometric analysis utilizing seven databases and a total of 478 articles that were selected based on the phrase “entrepreneurial mindset”. Data were extracted on 6 July 2024, and the bibliometric analysis consisted of four separate steps in the methodological approach. The findings identified six different clusters in which the entrepreneurial mindset adopted a process-oriented perspective, a concept that is underexplored in the current literature. The novelty in this study involves a cluster in the findings, labeled “the missing dynamics”, which warrants attention. Overall, the missing dynamics cluster in this bibliometric analysis offers originality and further research suggestions. By continuing to explore the process-oriented views of the entrepreneurial mindset, new value opportunities can be created, while the missing dynamics can be better understood. Full article

Multi-population mean-field game is a critical subclass of mean-field games (MFGs). It is a theoretically feasible multi-agent model for simulating and analyzing the game between multiple heterogeneous populations of interacting massive agents. Due to the factors of game complexity, dimensionality disaster and disturbances should be taken into account simultaneously to solve the multi-population high-dimensional stochastic MFG problem, which is a great challenge. We present CA-Net, a coupled alternating neural network approach for tractably solving multi-population high-dimensional MFGs. First, we provide a universal modeling framework for large-scale heterogeneous multi-agent game systems, which is strictly expressed as a multi-population MFG problem. Next, we generalize the potential variational primal–dual structure that MFGs exhibit, then phrase the multi-population MFG problem as a convex–concave saddle-point problem. Last but not least, we design a generative adversarial network (GAN) with multiple generators and multiple discriminators—the solving network—which parameterizes the value functions and the density functions of multiple populations by two sets of neural networks, respectively. In multi-group quadcopter trajectory-planning numerical experiments, the convergence results of HJB residuals, control, and average speed show the effectiveness of the CA-Net algorithm, and the comparison with baseline methods—cluster game, HJB-NN, Lax–Friedrichs, ML, and APAC-Net—shows the progressiveness of our solution method. Full article

A Domain Generation Algorithm (DGA) employs botnets to generate domain names through a communication link between the C&C server and the bots. A DGA can generate pseudo-random AGDs (algorithmically generated domains) regularly, a handy method for detecting bots on the C&C server. Unlike current DGA detection methods, AGDs can be identified with lightweight, promising technology. DGAs can prolong the life of a viral operation, improving its profitability. Recent research on the sensitivity of deep learning to various adversarial DGAs has sought to enhance DGA detection techniques. They have character- and word-level classifiers; hybrid-level classifiers may detect and classify AGDs generated by DGAs, significantly diminishing the effectiveness of DGA classifiers. This work introduces WordDGA, a hybrid RCNN-BiLSTM-based adversarial DGA with strong anti-detection capabilities based on NLP and cWGAN, which offers word- and hybrid-level evasion techniques. It initially models the semantic relationships between benign and DGA domains by constructing a prediction model with a hybrid RCNN-BiLSTM network. To optimize the similarity between benign and DGA domain names, it modifies phrases from each input domain using the prediction model to detect DGA family categorizations. The experimental results reveal that dodging numerous wordlists and mixed-level DGA classifiers with training and testing sets improves word repetition rate, domain collision rate, attack success rate, and detection rate, indicating the usefulness of cWGAN-based oversampling in the face of adversarial DGAs. Full article

The acquisition of deictic verbs is a significant milestone in language development. This complex process requires an understanding of the interplay between the personal pronouns “I/you” and deictic verbs. Although demonstrating the cognitive processes associated with deictic shifting through data is valuable, research issues regarding data accuracy and the spatial arrangement of the self and other remain unresolved. This pilot study aimed to quantitatively measure the body movements of Japanese-speaking children during their utterances of “come/go”. Twelve typically developing children aged 6–7 participated in this study. Multiple scenarios were set up where the researcher presented phrases using “come/go” with deictic gestures, such as moving one’s upper body forward or backward, and the participant replied with “come/go”. When performing a role, the researcher sat face-to-face or side-by-side with the participant, depending on the type of question–response. It is possible that there is a learning process whereby verbal responses using “come/go” align with corresponding body movements in the specific question type. This process is deeply involved in the development of perspective-taking abilities. Future research with relatively large samples and cross-cultural comparisons is warranted to deepen the understanding of this linguistic acquisition process and its implications. Full article

Recently, zero-shot image captioning (ZSIC) has gained significant attention, given its potential to describe unseen objects in images. This is important for real-world applications such as human–computer interaction, intelligent education, and service robots. However, the zero-shot image captioning method based on large-scale pretrained models may generate descriptions containing objects that are not present in the image, which is a phenomenon termed “object hallucination”. This is because large-scale models tend to predict words or phrases with high frequency, as seen in the training phase. Additionally, the method set a limitation to the description length, which often leads to an improper ending. In this paper, a novel approach is proposed to address and reduce the object hallucination and improper ending problem in the ZSIC task. We introduce additional emotion signals as guidance for sentence generation, and we find that proper emotion will filter words that do not appear in the image. Moreover, we propose a novel strategy that gradually extends the number of words in a sentence to confirm the generated sentence is properly completed. Experimental results show that the proposed method achieves the leading performance on unsupervised metrics. More importantly, the subjective examples illustrate the effect of our method in improving hallucination and generating properly ending sentences. Full article

(1) Background: “To change the world, we must first change the way the babies are being born”, said Michel Odent, the famous French obstetrician and pioneer of the “natural birth” movement. This quotable phrase has been widespread in activism campaigns, and it refers to a project for social change that goes beyond birth. Conceiving childbirth in the broader social context, it is not surprising that this emblematic quote inspires emancipatory struggles around birth. This paper results from a study of childbirth activism in different European contexts, where the author explores the emergence and modes of action of social movements advocating for the humanization of childbirth and women’s rights in pregnancy and childbirth. (2) Methods: Starting from the analysis of the main characteristics of childbirth activism, in this paper the author briefly analyzes the cases of organizations from Portugal, Spain and the Netherlands, as well as a campaign promoted by the European Network of Childbirth Associations (ENCA). The author mobilizes empirical data resulting from a triangulation approach, essentially based on documentary analysis, complemented by conversations and participant observation in different settings. (3) Results: Preliminary results show how childbirth activism is contributing to the construction of alternative conceptions of birth, challenging established paradigms. (4) Conclusion: In its differences and similarities, childbirth activism assumes distinct features, but it also has the ability to adapt and promote changes, depending on the specificities of the contexts where it operates. Full article

Weakly supervised temporal language grounding (TLG) aims to locate events in untrimmed videos based on natural language queries without temporal annotations, necessitating a deep understanding of semantic context across both video and text modalities. Existing methods often focus on simple correlations between query phrases and isolated video segments, neglecting the event-oriented semantic coherence and consistency required for accurate temporal grounding. This can lead to misleading results due to partial frame correlations. To address these limitations, we propose the Event-oriented State Alignment Network (ESAN), which constructs “start–event–end” semantic state sets for both textual and video data. ESAN employs relative entropy for cross-modal alignment through knowledge distillation from pre-trained large models, thereby enhancing semantic coherence within each modality and ensuring consistency across modalities. Our approach leverages vision–language models to extract static frame semantics and large language models to capture dynamic semantic changes, facilitating a more comprehensive understanding of events. Experiments conducted on two benchmark datasets demonstrate that ESAN significantly outperforms existing methods. By reducing false high correlations and improving the overall performance, our method effectively addresses the challenges posed by previous approaches. These advancements highlight the potential of ESAN to improve the precision and reliability of temporal language grounding tasks. Full article

This study focuses on accent shift or stylization to American English features in Anglophone pop-rock music and examines linguistic constraints alongside music-related considerations, as well as the effect of changes in musical genre on variable accent shift. The case study is the British band Mumford and Sons and their variable production of non-prevocalic rhotics as either present or absent. Mumford and Sons is of interest because they have displayed a change in their musical style throughout their career from Americana to alt-rock. The band’s four studio albums were auditorily analyzed and coded for rhotic vs. non-rhotic with aid from spectrograms. The linguistic factors considered were word class, preceding vowel according to the word’s lexical set, complexity of the preceding vowel, syllable complexity, stress, and location within the word and phrase. In addition, the effect of singing-related factors of syllable elongation and rhyming, and of the specific album, were also explored. Results show that rhoticity is favored in content words, stressed contexts, complex syllables, and NURSE words. This pattern is explained as stemming from the perceptual prominence of those contexts based on their acoustic and phonological characteristics. Results further show that syllable elongation leads to more rhoticity and that rhyming words tend to agree in their (non-)rhoticity. Finally, the degree of rhoticity decreases as the band departs from Americana in their later albums, highlighting the relevance of music genre for accent stylization. Full article

This article presents a comprehensive evaluation of traditional machine learning and deep learning models in analyzing sentiment trends within the SENT-COVID Twitter corpus, curated during the COVID-19 pandemic. The corpus, filtered by COVID-19 related keywords and manually annotated for polarity, is a pivotal resource for conducting sentiment analysis experiments. Our study investigates various approaches, including classic vector-based systems such as word2vec, doc2vec, and diverse phrase modeling techniques, alongside Spanish pre-trained BERT models. We assess the performance of readily available sentiment analysis libraries for Python users, including TextBlob, VADER, and Pysentimiento. Additionally, we implement and evaluate traditional classification algorithms such as Logistic Regression, Naive Bayes, Support Vector Machines, and simple neural networks like Multilayer Perceptron. Throughout the research, we explore different dimensionality reduction techniques. This methodology enables a precise comparison among classification methods, with BETO-uncased achieving the highest accuracy of 0.73 on the test set. Our findings underscore the efficacy and applicability of traditional machine learning and deep learning models in analyzing sentiment trends within the context of low-resource Spanish language scenarios and emerging topics like COVID-19. Full article

Background: The Chear open-set performance test (COPT), which uses a carrier phrase followed by a monosyllabic test word, is intended for clinical assessment of speech recognition, evaluation of hearing-device performance, and the fine-tuning of hearing devices for speakers of British English. This paper assesses practice effects, test–retest reliability, and the variability across lists of the COPT. Method: In experiment 1, 16 normal-hearing participants were tested using an initial version of the COPT, at three speech-to-noise ratios (SNRs). Experiment 2 used revised COPT lists, with items swapped between lists to reduce differences in difficulty across lists. In experiment 3, test–retest repeatability was assessed for stimuli presented in quiet, using 15 participants with sensorineural hearing loss. Results: After administration of a single practice list, no practice effects were evident. The critical difference between scores for two lists was about 2 words (out of 15) or 5 phonemes (out of 50). The mean estimated SNR required for 74% words correct was −0.56 dB, with a standard deviation across lists of 0.16 dB. For the participants with hearing loss tested in quiet, the critical difference between scores for two lists was about 3 words (out of 15) or 6 phonemes (out of 50). Full article