Search Results (79)

Forestry is recognized as one of the most dangerous industries in the world. To enhance forestry safety, autonomous machinery and safety systems for such machinery are essential. This study aims to introduce large language models (LLMs)—especially their extensions to images, vision–language models (VLMs)—to enable human-like decision-making for autonomous forest machinery. This research focused on VLMs as an active safety system that can adapt to environments and evaluated the effectiveness of a system that quantitatively makes decisions regarding hazard levels using contrastive language–image pretraining (CLIP). The results of industry type, tree state, and road state classification using pretrained models showed that for three tasks—forestry identification, hung-up tree detection, and road collapse sensing—the target classes consistently exhibited higher similarity with disaster texts compared with nontarget classes. Although the F1 scores were 0.693, 0.324 and 0.634, respectively—indicating that the system is insufficient as a direct active safety system—the application of a similarity threshold optimized to maintain a recall of 0.9 yielded F1 scores of 0.291 and 0.584 for tree state and road state, respectively. These results suggest that the system can potentially be used as a quantitative indicator of hazard by setting a threshold on the similarity score. Full article

Despite overwhelming scientific consensus that human activity drives climate change, public opinion in the United States remains sharply polarized along political lines. This project tests whether a theory-driven narrative intervention can reduce divergence between individuals skeptical of climate change and those who accept the scientific consensus. Guided by narrative transportation theory, we hypothesize that an inclusive, character-driven video grounded in the authentic language of skeptical audiences will reduce polarization and increase civic engagement. The study proceeds in three phases. Phase 1 uses focus group discussions to identify words, phrases, and perspectives used by skeptical and accepting participants. Phase 2 integrates these findings into the production of a 2–3 min narrative short film, refined through iterative audience testing. Phase 3 employs a stratified online experiment assessing climate attitudes, policy support, and activism behaviors before exposure, immediately after, and one week later. Mediators include narrative transportation, perceived similarity, and character identification. We test whether pre-exposure divergence narrows over time and whether engagement mechanisms explain observed changes. Findings will inform climate communication policy, intervention design, and broader research on depolarization in polarized public issues. Full article

Composed Video Retrieval (CoVR) enables flexible video search by retrieving a target video that reflects a specified modification to a query video. The triplet datasets—consisting of query videos, query text, and target videos—required for model training have been collected manually. Recent studies have explored automatic construction of training triplets for CoVR; however, most existing approaches rely heavily on caption similarity. This limitation is particularly problematic in soccer videos, where identical or highly similar captions can correspond to visually distinct situations, making it difficult to construct triplets with appropriate relationships. To address this issue, this paper proposes a multimodal triplet construction framework specialized for soccer videos. The key idea is to explicitly incorporate visual similarity alongside textual similarity. Specifically, candidate target videos are selected by combining visual similarity with commentary caption filtering, enabling the identification of videos that are visually similar yet semantically different. The semantic difference between videos is then generated as query text using a large language model (LLM) without manual annotation. Furthermore, a multimodal large language model (MLLM) is introduced to estimate whether the generated modification is visually and semantically consistent with the video pair. Rather than replacing human verification, this step provides an automated screening signal to identify potentially unreliable triplets. The experiments show that the proposed framework automatically constructs triplets with reasonable validity under limited human validation. These results demonstrate the potential of scalable triplet construction for CoVR in soccer videos. Full article

Component obsolescence and supply-chain disruptions increasingly force engineers to spend significant time manually searching and comparing PDF datasheets to identify compatible replacement parts. We propose an AI-powered datasheet assistant based on a Retrieval-Augmented Generation (RAG) pipeline that automatically processes datasheets to accelerate component identification and matching. The core contribution is a summary-driven retrieval mechanism: a Large Language Model (LLM) generates a structured semantic summary of an input datasheet, and the vector embedding of this summary is used to retrieve semantically similar components from a reference database. The system also supports natural language question answering and structured component comparison. Its architecture separates scalable text-only reference indexing from more expensive query-time summarization and reranking. Validation includes a controlled synthetic benchmark and a pilot-scale real-world evaluation on 18 publicly listed microcontroller datasheets grouped into six engineering families. The synthetic benchmark is used to assess pipeline behavior under controlled conditions, while the real-world evaluation measures performance on heterogeneous manufacturer datasheets. In the real-world evaluation, structured summaries generated with Claude Sonnet 4.5 combined with cross-encoder reranking achieved a 72.2% Family Retrieval Rate at $k=1$ (13/18; Wilson 95% CI: 49.1–87.5%). Additional experiments with local LLM summaries indicate that retrieval performance depends strongly on summary quality and model capability, with lightweight local summarizers producing lower first-candidate retrieval performance in this setup. The analysis further reports confidence intervals, no-summary baselines, chunking sensitivity, and an Image Reference Rate metric used as a lexical reference proxy rather than a direct measure of visual grounding. Full article

Frequent accidents in the construction sector arise from the dynamic coupling of multiple risk factors, while conventional single-factor approaches fail to capture the underlying complexity. Drawing on 702 accident investigation reports, this study develops an intelligent, data-driven framework that integrates large language model–based risk identification with association rule mining to systematically uncover risk factors and their coupling patterns. A DeepSeek-based model is employed to extract risk factors from unstructured text, followed by cosine similarity–based optimization to refine factor representations. The FP-Growth algorithm is then applied to identify strong association rules among risk factors. The results reveal that deficiencies in the management dimension account for 68.30% of all identified risks, with inadequate safety education and training emerging as the central hub in the risk coupling network, which is further corroborated by complex network analysis. Moreover, a cascading transmission pathway is identified, whereby environmental deficiencies induce weakened safety awareness, which in turn leads to unsafe behaviors. These findings further demonstrate the nonlinear amplification effects arising from concurrent management failures. By establishing a transformation pathway from unstructured textual data to structured risk knowledge, this study provides a robust, data-driven foundation for precise risk identification and systematic prevention in construction safety management. Full article

Artificial intelligence has attracted increasing attention in the construction industry; however, automated time scheduling remains limited in practical applications. Schedule development remains manual, requiring planners to analyze project documents, define activities, estimate durations, and identify relationships based on logical sequence. This process primarily depends on individual experience and skills, making it both time-consuming and prone to human error. From an engineering design perspective, delayed or inconsistent schedule development weakens design-to-construction feedback, limiting the ability to evaluate constructability and time implications of alternative design decisions during early-stage planning. This study proposes an integrated BIM–Natural Language Processing (NLP) framework to automate activity identification, duration estimation, and logical sequencing for construction scheduling. The framework extracts project data from Revit, organizes it into a bill of quantities format, and then generates an activity list, each activity with a unique ID. Using Sentence-BERT (SBERT) embeddings, the framework estimates activity durations based on semantic similarity. The same semantic process is combined with rule-based reasoning to identify logical relationships, including sequences, supported by an Excel-based reference dictionary that includes logical relationships, productivity, and ID structure. Finally, the framework incorporates a crashing module that proportionally adjusts the duration of activities on the longest path to target the project’s completion time without violating relationships. The proposed framework was validated using real construction project data and produced reliable results. By producing a tool-ready schedule directly from design-model information, the proposed workflow enables earlier schedule feedback loops and supports design-informed planning by allowing designers and planners to assess the time consequences of model-driven scope changes. The results demonstrate that integrating BIM and NLP can transform conventional schedules into faster, more consistent processes, thereby supporting the construction industry. Full article

Developing reliable Arabic question answering (QA) systems for Islamic fatwas requires datasets that capture the linguistic complexity and multi-step reasoning inherent in jurisprudential inquiries. However, the existing Arabic religious QA datasets primarily focus on direct retrieval or classification, often failing to address the multi-hop reasoning necessary for complex fatwa questions. To bridge this gap, we introduce MAFQA, a benchmark dataset specifically designed for multi-hop Arabic fatwa question answering. MAFQA was constructed from an extensive corpus of authentic fatwa records sourced from authoritative Islamic institutions. The dataset was developed via a semi-automated pipeline that integrates expert-guided identification of complex inquiries with a structured decomposition framework. This framework employs automated reasoning-pattern classification, semantic feature extraction, and template-guided annotation of subquestions and subanswers, followed by rigorous validation to ensure contextual grounding, logical coherence, and structural consistency. To evaluate the utility of the dataset, we conduct an extensive benchmarking study using Arabic-specialized, multilingual, and instruction-tuned language models across two primary tasks: question decomposition (QD) and generative question answering (QA). Performance is assessed using a comprehensive suite of lexical, semantic, relevance, and faithfulness metrics. Experimental results demonstrate that Arabic-specialized models consistently outperform their multilingual counterparts, with AraT5-base and AraBART achieving the highest performance in terms of lexical similarity, semantic alignment, and answer faithfulness. Full article

Selecting a source language for zero-shot cross-lingual transfer is typically done by intuition or by defaulting to English, despite large performance differences across language pairs. We study whether linguistic similarity can predict transfer performance and support principled source-language selection. We introduce quantified WALS (qWALS), a typology-based similarity metric derived from features in the World Atlas of Language Structures, and evaluate it against existing similarity baselines. Validation uses three complementary signals: computational similarity scores, zero-shot transfer performance of multilingual transformers (mBERT and XLM-R) on four NLP tasks (dependency parsing, named entity recognition, sentiment analysis, and abusive language identification) across eight languages, and an expert-linguist similarity survey. Across tasks and models, higher linguistic similarity is associated with better transfer, and the survey provides independent support for the computational metrics. Full article

Backdoor attacks pose a critical security risk to pre-trained language models (PLMs) by utilizing concealed triggers to manipulate model outputs. Existing defense strategies largely depend on statistical thresholds, which often struggle to identify sophisticated backdoor samples that exhibit high cognitive similarity to benign data. Such similarities make precise threshold calibration difficult, frequently leading to unreliable or failed detection. To overcome these limitations, we propose a backdoor detection method based on consensus deviation, shifting the defensive paradigm from surface-level statistical metrics to deep cognitive consensus verification. This approach obviates the reliance on fixed thresholds, enabling the more robust identification of covert triggers. Extensive experiments on the SST-2, HSOL, and AG‘s News datasets revealed that our method achieved significantly lower attack success rates (ASRs) and enhanced robustness compared with the current baselines across word-, sentence-, and structural-level attack scenarios. Full article

Genomic Language Models (GLMs), leveraging their vast parameter scales and the similarities between DNA sequences and natural languages, demonstrate immense potential in processing large-scale genomic data and elucidating gene regulation and evolutionary relationships. However, the cross-species generalization capability of large genomic models has not yet been systematically evaluated. This study addresses this critical gap by benchmarking five GLMs (DNABERT-2, GROVER, HyenaDNA, NT-V2, and AgroNT) and a CNN baseline model using human (Homo sapiens) and rice (Oryza sativa) genomes across four downstream tasks: promoter detection, transcription start site (TSS) scanning, species classification, and gene region identification, through both zero-shot testing and fine-tuning. During testing, factors such as hyperparameters, early stopping protocols, and computational resources were fixed to ensure fairness, enabling us to systematically evaluate their performance and cross-species generalization capabilities. The results were further analyzed from multiple mathematical and representational perspectives to provide a more rigorous and objective assessment of each model’s performance. The results show that AgroNT consistently leads on rice tasks, while NT-V2 and DNABERT-2 achieved the best overall performance in fine-tuning and zero-shot experiments, respectively. Although their pretraining data did not include plants, they demonstrate excellent performance on rice-related tasks thanks to cross-species pretraining that enhances their generalization ability across human–rice domains. This benchmark study offers guidance on selecting appropriate genomic language models based on task characteristics and provides insights for future development in this field. Full article

Manual pill identification is often inefficient and error-prone due to the large variety of medications and frequent visual similarity among pills, leading to misuse or dispensing errors. These challenges are exacerbated when pill imprints are engraved, curved, or irregularly arranged, conditions under which conventional optical character recognition (OCR)-based methods degrade significantly. To address this problem, we propose GO-PILL, a geometry-aware OCR pipeline for robust pill imprint recognition. The framework extracts text centerlines and imprint regions using the TextSnake algorithm. During imprint refinement, background noise is suppressed and contrast is enhanced to improve the visibility of embossed and debossed imprints. The imprint localization and alignment stage then rectifies curved or obliquely oriented text into a linear representation, producing geometrically normalized inputs suitable for OCR decoding. The refined imprints are processed by a multimodal OCR module that integrates a non-autoregressive language–vision fusion architecture for accurate character-level recognition. Experiments on a pill image dataset from the U.S. National Library of Medicine show that GO-PILL achieves an F1-score of 81.83% under set-based evaluation and a Top-10 pill identification accuracy of 76.52% in a simulated clinical scenario. GO-PILL consistently outperforms existing methods under challenging imprint conditions, demonstrating strong robustness and practical feasibility. Full article

A Smart campus increasingly operates on the basis of data-driven operations, but an increasing demand for energy puts their control over costs and sustainability at risk. This study addresses the challenge of anticipating and managing energy consumption peaks in multi-campus environments by proposing a hybrid framework that combines advanced time-series forecasting models with a large language model (LLM)-driven multi-agent system. Based on the UNICON dataset, LSTM, CNN, GRU, and a combination architecture are trained and compared in terms of MAE and RMSE. The hybrid configuration achieves the greatest forecasting results by returning the minimum loss values. For the identification of critical periods, we employed a strategy based on median thresholding, which offers a categorization into low, normal, and extreme category, allowing the targeting of peak mitigation actions. We also introduce a multi-agent system based on the LLM, including the data aggregator, the forecaster, and the policy advisor, which create actionable policies informed by context. We also compare LLMs (Qwen-2.5, Gemma-2, Phi-4, Mistral, Llama-3.3) in terms of context accuracy, response relevance, semantic similarity, and retrieval/recall accuracy and fidelity, with Llama-3.3 achieving the best overall results. This framework has shown great potential, not only for energy consumption forecasting but also for developing precise policies on how to effectively manage energy consumption peaks. Full article

In the context of remanufacturing, particularly mobile device refurbishing, effective operator training is crucial for accurate defect identification and process inspection efficiency. This study examines the application of Natural Language Processing (NLP) techniques to evaluate operator expertise based on subjective textual responses gathered during a defect analysis task. Operators were asked to describe screen defects using open-ended questions, and their responses were compared with expert responses to evaluate their accuracy and consistency. We employed four NLP models, including finetuned Sentence-BERT (SBERT), pre-trained SBERT, Word2Vec, and Dice similarity, to determine their effectiveness in interpreting short, domain-specific text. A novel disagreement tagging framework was introduced to supplement traditional similarity metrics with explainable insights. This framework identifies the root causes of model–human misalignment across four categories: defect type, severity, terminology, and location. Results show that a finetuned SBERT model significantly outperforms other models by achieving Pearsons’s correlation of 0.93 with MAE and RMSE scores of 0.07 and 0.12, respectively, providing more accurate and context-aware evaluations. In contrast, other models exhibit limitations in semantic understanding and consistency. The results highlight the importance of finetuning NLP models for domain-specific applications and demonstrate how qualitative tagging methods can enhance interpretability and model debugging. This combined approach indicates a scalable and transparent methodology for the evaluation of operator responses, supporting the development of more effective training programmes in industrial settings where remanufacturing and sustainability generally are a key performance metric. Full article

Traditional methods for patient cohort identification from Electronic Health Records (EHRs) are often slow, labor-intensive, and fail to capture the rich semantic nuance embedded in unstructured clinical narratives. This paper introduces a scalable, end-to-end framework that creates a synergy between deep medical embeddings and Approximate Nearest Neighbor Search (ANNs) to overcome these limitations. We detail a complete pipeline that begins with preprocessing multi-modal EHR data and creating holistic patient representations using a domain-specific language model combined with an intelligent gated fusion mechanism. These high-dimensional embeddings are then indexed using an ANN method to enable near real-time retrieval. A comprehensive experimental evaluation was conducted on the MIMIC-III and MIMIC-IV datasets, comparing the retrieval performance of ClinicalBERT against BioBERT across several ANN algorithms. The results demonstrate that the combination of ClinicalBERT and HNSW consistently achieves the highest retrieval accuracy, with F1-Scores exceeding 0.78, and query latencies under 10 ms. This framework enables a paradigm shift towards high-speed, semantic patient similarity search, with significant implications for accelerating clinical trial recruitment, augmenting clinical decision support, and paving the way for a new era in data-driven precision medicine. Full article

Aberrant protein–protein interactions (PPIs) underpin a plethora of human diseases, and disruption of these harmful interactions constitute a compelling treatment avenue. Advances in computational approaches to PPI prediction have closely followed progress in deep learning and natural language processing. In this review, we outline the state-of-the-art methods for sequence-based PPI prediction and explore their impact on target identification and drug discovery. We begin with an overview of commonly used training data sources and techniques used to curate these data to enhance the quality of the training set. Subsequently, we survey various PPI predictor types, including traditional similarity-based approaches, and deep learning-based approaches with a particular emphasis on transformer architecture. Finally, we provide examples of PPI prediction in system-level proteomics analyses, target identification, and designs of therapeutic peptides and antibodies. This review sheds light on sequence-based PPI prediction, a broadly applicable alternative to structure-based methods, from a unique perspective that emphasizes their roles in the drug discovery process and rigorous model assessment. Full article