Search Results (23)

While handwritten signatures are a staple of biometric authentication, conventional verification models typically rely on Euclidean space assumptions, restricting the capture of complex, intrinsic signature structures. To address this, offline signature verification has increasingly modeled signatures as points on the Symmetric Positive Definite (SPD) manifold. Nevertheless, selecting an appropriate metric on this manifold for a given problem remains a significant challenge, typically relying on heuristic trial-and-error processes. To solve this, our primary contribution is a novel, end-to-end Riemannian framework featuring the Collective Background Knowledge (CBK) mechanism. CBK establishes synthetic writers as Riemannian centers, utilizing a learnable $\alpha \beta$-Log-Determinant divergence to adaptively discover the optimal local geometry from data. Instead of computing the direct distance between two SPD signature representations, we evaluate them relationally by measuring how each signature diverges from the shared CBK reference centers. These individual deviations form unique relational profiles for each signature, which are then compared using the dichotomy transform to create a dissimilarity vector. By jointly optimizing the CBK parameters under an SPD metric-learning approach, our model separates effectively similar and dissimilar pairs of signatures. Evaluated across five datasets under challenging blind intra- and cross-lingual conditions, our geometry-aware framework demonstrates robust generalization and competitive performance. Full article

Speech-based non-invasive screening offers a cost-effective and scalable approach for the early detection of Alzheimer’s disease (AD). However, the clinical utility of deep learning models remains severely constrained by the scarcity of labeled speech data in low-resource languages, necessitating cross-lingual transfer learning. Conventional domain adaptation paradigms typically assume semantically consistent feature domains and focus heavily on aligning marginal distributions; however, they suffer catastrophic performance degradation when applied to cross-lingual pathologic speech. By analyzing disease-associated representation vectors within a self-supervised HuBERT space, we uncover a systematic mechanism driving this failure, a phenomenon we term cross-lingual polarity flip, where the direction of disease-relative-to-control feature offsets fundamentally reverses between languages. While prior multilingual studies have largely discarded such dimensional inconsistencies as ungeneralizable noise, a 500-round Monte Carlo stability analysis demonstrates that these flips occur in a highly stable, structural manner across 18.3% of top discriminative dimensions. Leveraging this insight, we introduce Monte Carlo Polarity Flip Calibration (MC-PFC), a few-shot framework designed to explicitly rectify flip orientations. Requiring only five labeled support samples per class from the target domain, MC-PFC robustly estimates direction flips via a separability-weighted ensemble voting mechanism. Evaluated on a strictly held-out Chinese blind test set, MC-PFC achieves an area under the receiver operating characteristic curve (AUC) of 0.871, recovering 99.5% of the performance achieved by a full in-domain trained upper bound (AUC = 0.875). Ablation experiments confirm that direction calibration yields a substantial +0.361 AUC gain, vastly outperforming standard distribution alignment (+0.081). This work establishes a data-efficient paradigm for cross-lingual medical analysis, shifting the clinical AI focus from discarding cross-lingual discrepancies to actively modeling and calibrating them. Full article

Large language models (LLMs) are increasingly adopted for Verilog code generation, yet existing benchmarks assume English-only prompts, overlooking the linguistic diversity of the global FPGA engineering community. We introduce Multi-VerilogEval, the first multilingual Verilog benchmark, built from 156 unique underlying tasks instantiated in four languages (English, Japanese, Hindi, and Mongolian), yielding 624 language-specific test cases. Our evaluation of four representative LLMs reveals a silent failure pattern: syntactic correctness remains high (∼90%) across languages, but functional correctness degrades by up to 23.9% for non-English prompts in open-source and domain-specific models, while commercial models remain near-parity. Hidden-state analysis suggests that multilingual bias is associated with persistent cross-lingual representation divergence throughout the network, which becomes most pronounced in the final layers that directly drive token generation. As fine-tuning and common prompt-based mitigations remain impractical or unreliable for multilingual RTL, we propose HE-ICL (Hard-Example In-Context Learning), a train-free method that constructs few-shot hard-example demonstrations from cross-lingually difficult cases. HE-ICL closes 80–100% of the multilingual gap without any parameter updates, achieving near-parity with or exceeding the English reference level across all evaluated HE-ICL settings. Full article

Integrating structured linguistic resources into deep learning architectures represents a key challenge in domain-oriented NLP. This study proposes a framework for incorporating knowledge from a military thesaurus of the Ground Forces, structured according to the XML Zthes standard, into pre-trained transformed language models, including KazBERT, multilingual BERT, and XLM-RoBERTA. The approach addresses two interrelated tasks in specialized terminology processing: concept linking and semantic search. Unlike existing knowledge-injection methods designed primarily for general-domain applications, this framework formalizes the mapping of Zthes elements, such as Term, Broader term, Narrower term, Related term, ScopeNote, Language, and Source into structured textual representations that can be directly processed by transformer architectures. Fine-tuning is conducted on a dataset of 18,400 training instances automatically generated from the thesaurus, including synonym pairs, hierarchical relations (hyperonymy and hyponymy), associative links, and definitional descriptions. Experimental evaluation demonstrated that thesaurus-enriched models outperform baseline architectures across all major metrics. XLM-RoBERTA model achieves F1 = 0.84 and Top-5 accuracy = 0.94 in the concept linking task, representing a five-point improvement over the baseline. The model reaches Macro-F1 = 0.84 across four relation types. Results obtained on a specialized test set derived from terminology databases of Kazakhstan’s Armed Forces confirm robust cross-lingual generalization across Kazakh, Russian and English military discourse. Full article

Against the background of highly interconnected global capital markets and rapidly propagating cross-lingual information streams, traditional anomaly detection paradigms based solely on single-modality numerical time-series sensors are insufficient for forward-looking risk sensing. From the perspective of artificial intelligence-driven sensing, this study proposes a multilingual semantic–numerical collaborative Transformer framework to construct a unified multimodal financial sensing architecture for intelligent anomaly sensing and risk perception. Within the proposed sensing paradigm, multilingual texts are conceptualized as semantic sensors that continuously emit event-driven sensing signals, while market prices, trading volumes, and order book dynamics are modeled as heterogeneous numerical sensor streams reflecting behavioral market sensing responses. These heterogeneous sensors are jointly integrated through a cross-modal sensor fusion architecture. A cross-modal temporal alignment attention mechanism is designed to explicitly model dynamic lag structures between semantic sensing signals and numerical sensor responses, enabling temporally adaptive sensor-level alignment and fusion. To enhance sensing robustness, a multilingual semantic noise-robust encoding module is introduced to suppress unreliable textual sensor noise and stabilize cross-lingual semantic sensing representations. Furthermore, a semantic–numerical collaborative risk fusion module is constructed within a shared latent sensing space to achieve adaptive sensor contribution weighting and cross-sensor feature coupling, thereby improving anomaly sensing accuracy and robustness under complex multimodal sensing environments. Extensive experiments conducted on real-world multi-market financial sensing datasets demonstrate that the proposed artificial intelligence-driven sensing framework significantly outperforms representative statistical and deep learning baselines. The framework achieves a Precision of $0.852$, Recall of $0.781$, F1-score of $0.815$, and an AUC of $0.892$, while substantially improving early warning time in practical risk sensing scenarios. In cross-market transfer settings, the proposed sensing architecture maintains stable anomaly sensing performance under bidirectional domain shifts, with AUC consistently exceeding $0.86$, indicating strong structural generalization across heterogeneous sensing environments. Ablation analysis further verifies that temporal sensor alignment, semantic sensor denoising, and collaborative cross-sensor risk coupling contribute independently and synergistically to the overall sensing performance. Overall, this study establishes a scalable multimodal intelligent sensing framework for dynamic financial anomaly sensing, providing an effective artificial intelligence-driven sensing solution for cross-market risk surveillance and adaptive financial signal sensing. Full article

Stance detection has emerged as an essential tool in natural language processing for understanding how individuals express agreement, disagreement, or neutrality toward specific targets in social and online discourse. It plays a crucial role in bilingual and multilingual environments, including English-Arabic social media ecosystems, where differences in language structure, discourse style, and data availability pose significant challenges for reliable stance modelling. Existing approaches often struggle with target awareness, cross-lingual generalization, robustness to noisy user-generated text, and the interpretability of model decisions. This study aims to build a reliable, explainable target-aware bilingual stance-detection framework that generalizes across heterogeneous stance formats and languages without retraining on a dataset specific to the target language. Thus, a unified dual-encoder architecture based on mDeBERTa-v3 is proposed. Cross-language contrastive learning offers an auxiliary training objective to align English and Arabic stance representations in a common semantic space. Robustness-oriented regularization is used to mitigate the effects of informal language, vocabulary variation, and adversarial noise. To promote transparency and trustworthiness, the framework incorporates token-level rationale extraction, enables fine-grained interpretability, and supports analysis of hallucination. The proposed model is tested on a combined bilingual test set and two structurally distinct zero-shot benchmarks: MT-CSD and AraStance. Experimental results show consistent performance, with accuracies of 85.0% and 86.8% and F1-scores of 84.7% and 86.8% on the zero-shot benchmarks, confirming stable performance and realistic generalization. Ultimately, these findings reveal that effective bilingual stance detection can be achieved via explicit target conditioning, cross-lingual alignment, and explainability-driven design. Full article

Advancements in artificial intelligence have significantly enhanced communication for individuals with hearing impairments. This study presents a robust cross-lingual Sign Language Recognition (SLR) framework for Turkish, American English, and Arabic sign languages. The system utilizes the lightweight MediaPipe library for efficient hand landmark extraction, ensuring stable and consistent feature representation across diverse linguistic contexts. Datasets were meticulously constructed from nine public-domain sources (four Arabic, three American, and two Turkish). The final training data comprises curated image datasets, with frames for each language carefully selected from varying angles and distances to ensure high diversity. A comprehensive comparative evaluation was conducted across three state-of-the-art deep learning architectures—ConvNeXt (CNN-based), Swin Transformer (ViT-based), and Vision Mamba (SSM-based)—all applied to identical feature sets. The evaluation demonstrates the superior performance of contemporary vision Transformers and state space models in capturing subtle spatial cues across diverse sign languages. Our approach provides a comparative analysis of model generalization capabilities across three distinct sign languages, offering valuable insights for model selection in pose-based SLR systems. Full article

Deep anomaly detection in multilingual environments remains challenging due to limited labeled data, semantic inconsistency across languages, and the unstable distribution of rare abnormal patterns. These challenges are particularly severe in low-resource scenarios—characterized by scarce labeled anomaly data and non-standardized terminology—where conventional supervised or transfer-based models suffer from semantic drift and feature mismatch. To address these limitations, a data-driven cross-lingual anomaly detection framework, LR-SSAD, is proposed. Targeting paired text and behavioral data without requiring parallel translation corpora, the framework is built upon the joint optimization of complementary self-supervised objectives. A cross-lingual masked prediction module is designed to capture language-invariant semantic structures to align semantic spaces, while a Mamba-based sequence reconstruction module leverages its linear computational complexity ($\mathcal{O}\left(N\right)$) to efficiently model long-range dependencies in transaction histories, overcoming the computational bottlenecks of quadratic attention mechanisms. To further enhance robustness under noisy supervision, a noise-aware pseudo-label refinement mechanism is introduced. Evaluated on a newly constructed real-world financial dataset (spanning January–June 2023) comprising 1.2 million multilingual texts and 420,000 transaction sequences, experimental results demonstrate that LR-SSAD achieves substantial improvements over state-of-the-art baselines. The model achieves an accuracy of 0.932, a precision of 0.914, a recall of 0.891, and an F1-score of 0.902, with the Area Under the Curve (AUC) reaching 0.948. The proposed framework provides a scalable and data-efficient solution for anomaly detection in real-world multilingual environments. Full article

Cyberbullying has become a pervasive threat on social media, impacting the safety and wellbeing of users worldwide. Most existing studies focus on monolingual content, limiting their applicability to online environments. This study aims to develop an approach that accurately detects abusive content in bilingual settings. Given the large volume of online content in English and Arabic, we propose a bilingual cyberbullying detection approach designed to deliver efficient, scalable, and robust performance. Several datasets were combined, processed, and augmented before proposing a cyberbullying identification approach. The proposed model (FusionBullyNet) is based on fine-tuning of two transformer models (RoBERTa-base + bert-base-arabertv02-twitter), attention-based fusion, gradually unfreezing the layers, and label smoothing to enhance generalization. The test accuracy of 0.86, F1 scores of 0.83 for bullying and 0.88 for no bullying, and an overall ROC-AUC of 0.929 were achieved with the proposed approach. To assess the robustness of the proposed models, several multilingual models, such as XLM-RoBERTa-Base, Microsoft/mdeberta-v3-base, and google-bert/bert-base-multilingual-cased, were also trained in this study, and all achieved a test accuracy of 0.84. Furthermore, several machine learning models were trained in this study, and Logistic Regression, XGBoost Classifier, and Light GBM Classifier achieved the highest accuracy of 0.82. These results demonstrate that the proposed approach provides a reliable, high-performance solution for cyberbullying detection, contributing to safer online communication environments. Full article

The pursuit of universal, symmetric semantic representations within large language models (LLMs) faces a fundamental challenge: the inherent asymmetry of natural languages. Different languages exhibit vast disparities in syntactic structures, lexical choices, and cultural nuances, making the creation of a truly shared, symmetric embedding space a non-trivial task. This paper aims to address this critical problem by introducing a novel framework to forge robust and symmetric multilingual sentence embeddings. Our approach, named DACL (Dynamic Asymmetric Contrastive Learning), is anchored in two powerful asymmetric learning paradigms: Contrastive Learning and Dynamic Curriculum Learning (DCL). We extend Contrastive Learning to the multilingual context, where it asymmetrically treats semantically equivalent sentences from different languages (positive pairs) and sentences with distinct meanings (negative pairs) to enforce semantic symmetry in the target embedding space. To further refine this process, we incorporate Dynamic Curriculum Learning, which introduces a second layer of asymmetry by dynamically scheduling training instances from easy to hard. This dual-asymmetric strategy enables the model to progressively master complex cross-lingual relationships, starting with more obvious semantic equivalences and advancing to subtler ones. Our comprehensive experiments on benchmark cross-lingual tasks, including sentence retrieval and cross-lingual classification (XNLI, PAWS-X, MLDoc, MARC), demonstrate that DACL significantly outperforms a wide range of established baselines. The results validate our dual-asymmetric framework as a highly effective approach for forging robust multilingual embeddings, particularly excelling in tasks involving complex linguistic asymmetries. Ultimately, this work contributes a novel dual-asymmetric learning framework that effectively leverages linguistic asymmetry to achieve robust semantic symmetry across languages. It offers valuable insights for developing more capable, fair, and interpretable multilingual LLMs, emphasizing that deliberately leveraging asymmetry in the learning process is a highly effective strategy. Full article

Under the growing demand for processing multimodal and cross-lingual information, traditional retrieval systems have encountered substantial limitations when handling heterogeneous inputs such as images, textual layouts, and multilingual language expressions. To address these challenges, a unified retrieval framework has been proposed, which integrates visual features from images, layout-aware optical character recognition (OCR) text, and bilingual semantic representations in Chinese and English. This framework aims to construct a shared semantic embedding space that mitigates semantic discrepancies across modalities and resolves inconsistencies in cross-lingual mappings. The architecture incorporates three main components: a visual encoder, a structure-aware OCR module, and a multilingual Transformer. Furthermore, a joint contrastive learning loss has been introduced to enhance alignment across both modalities and languages. The proposed method has been evaluated on three core tasks: a single-modality retrieval task from image → OCR, a cross-lingual retrieval task between Chinese and English, and a joint multimodal retrieval task involving image, OCR, and language inputs. Experimental results demonstrate that, in the joint multimodal setting, the proposed model achieved a Precision@10 of 0.693, Recall@10 of 0.684, nDCG@10 of 0.672, and F1@10 of 0.685, substantially outperforming established baselines such as CLIP, LayoutLMv3, and UNITER. Ablation studies revealed that removing either the structure-aware OCR module or the cross-lingual alignment mechanism resulted in a decrease in mean reciprocal rank (MRR) to 0.561, thereby confirming the critical role of these components in reinforcing semantic consistency across modalities. This study highlights the powerful potential of large language models in multimodal semantic fusion and retrieval tasks, providing robust solutions for large-scale semantic understanding and application scenarios in multilingual and multimodal contexts. Full article

Although Singing Voice Synthesis (SVS) has revolutionized audio content creation, global linguistic diversity remains challenging. Current SVS research shows scant exploration of cross-lingual generalization, as fragmented, language-specific phoneme encodings (e.g., Pinyin, ARPA) hinder unified phonetic modeling. To address this challenge, we built a four-language dataset based on GTSinger’s speech data, using the International Phonetic Alphabet (IPA) for consistent phonetic representation and applying precise segmentation and calibration for improved quality. In particular, we propose a novel method of decomposing IPA phonemes into letters and diacritics, enabling the model to deeply learn the underlying rules of pronunciation and achieve better generalization. A dynamic IPA adaptation strategy further enables the application of learned phonetic representations to unseen languages. Based on VISinger2, we introduce Transinger, an innovative cross-lingual synthesis framework. Transinger achieves breakthroughs in phoneme representation learning by precisely modeling pronunciation, which effectively enables compositional generalization to unseen languages. It also integrates Conformer and RVQ techniques to optimize information extraction and generation, achieving outstanding cross-lingual synthesis performance. Objective and subjective experiments have confirmed that Transinger significantly outperforms state-of-the-art singing synthesis methods in terms of cross-lingual generalization. These results demonstrate that multilingual aligned representations can markedly enhance model learning efficacy and robustness, even for languages not seen during training. Moreover, the integration of a strategy that splits IPA phonemes into letters and diacritics allows the model to learn pronunciation more effectively, resulting in a qualitative improvement in generalization. Full article

Diacritical Arabic (DA) refers to Arabic text with diacritical marks that guide pronunciation and clarify meanings, making their recognition crucial for accurate linguistic interpretation. These diacritical marks (short vowels) significantly influence meaning and pronunciation, and their accurate recognition is vital for the effectiveness of automatic speech recognition (ASR) systems, particularly in applications requiring high semantic precision, such as voice-enabled translation services. Despite its importance, leveraging advanced machine learning techniques to enhance ASR for diacritical Arabic has remained underexplored. A key challenge in developing DA ASR is the limited availability of training data. This study introduces a transformer-based approach leveraging transfer learning and data augmentation to address these challenges. Using a cross-lingual speech representation (XLSR) model pretrained on 53 languages, we fine-tune it on DA and integrate connectionist temporal classification (CTC) with transformers for improved performance. Data augmentation techniques, including volume adjustment, pitch shift, speed alteration, and hybrid strategies, further mitigate data limitations, significantly reducing word error rates (WER). Our methods achieve a WER of 12.17%, outperforming traditional ASR systems and setting a new benchmark for DA ASR. These findings demonstrate the potential of advanced machine learning to address longstanding challenges in DA ASR and enhance its accuracy. Full article

Social media serves as a platform for netizens to stay informed and express their opinions through the Internet. Currently, the social media discourse environment faces a significant security threat—offensive comments. A group of users posts comments that are provocative, discriminatory, and objectionable, intending to disrupt online discussions, provoke others, and incite intergroup conflict. These comments undermine citizens’ legitimate rights, disrupt social order, and may even lead to real-world violent incidents. However, current automatic detection of offensive language primarily focuses on a few high-resource languages, leaving low-resource languages, such as Malay, with insufficient annotated corpora for effective detection. To address this, we propose a zero-shot, cross-language unsupervised offensive language detection (OLD) method using a dual-branch mBERT transfer approach. Firstly, using the multi-language BERT (mBERT) model as the foundational language model, the first network branch automatically extracts features from both source and target domain data. Subsequently, Sinkhorn distance is employed to measure the discrepancy between the source and target language feature representations. By estimating the Sinkhorn distance between the labeled source language (e.g., English) and the unlabeled target language (e.g., Malay) feature representations, the method minimizes the Sinkhorn distance adversarially to provide more stable gradients, thereby extracting effective domain-shared features. Finally, offensive pivot words from the source and target language training sets are identified. These pivot words are then removed from the training data in a second network branch, which employs the same architecture. This process constructs an auxiliary OLD task. By concealing offensive pivot words in the training data, the model reduces overfitting and enhances robustness to the target language. In the end-to-end framework training, the combination of cross-lingual shared features and independent features culminates in unsupervised detection of offensive speech in the target language. The experimental results demonstrate that employing cross-language model transfer learning can achieve unsupervised detection of offensive content in low-resource languages. The number of labeled samples in the source language is positively correlated with transfer performance, and a greater similarity between the source and target languages leads to better transfer effects. The proposed method achieves the best performance in OLD on the Malay dataset, achieving an F1 score of 80.7%. It accurately identifies features of offensive speech, such as sarcasm, mockery, and implicit expressions, and showcases strong generalization and excellent stability across different target languages. Full article

Unlike previous approaches in speech emotion recognition (SER), which typically extract emotion embeddings from a trained classifier consisting of fully connected layers and training data without considering contextual information, this research introduces a novel approach. It integrates contextual information into the feature extraction process. The proposed approach is based on the WavLM representation and incorporates a contextual transform, along with fully connected layers, training data, and corresponding label information, to extract single-lingual WavLM domain emotion embeddings (SL-WDEEs) and cross-lingual WavLM domain emotion embeddings (CL-WDEEs) for single-lingual and cross-lingual SER, respectively. To extract CL-WDEEs, multi-task learning is employed to remove language information, marking it as the first work to extract emotion embeddings for cross-lingual SER. Experimental results on the IEMOCAP database demonstrate that the proposed SL-WDEE outperforms some commonly used features and known systems, while results on the ESD database indicate that the proposed CL-WDEE effectively recognizes cross-lingual emotions and outperforms many commonly used features. Full article