Search Results (56)

In today’s consumer-oriented market environment, enterprises face increasingly intense competition, making the identification of key factors in demand-driven product design critically important. Identifying critical product design elements facilitates the optimal allocation of design resources and enhances product competitiveness to better meet user expectations. This study proposes an attribute-network-based method for identifying key product design elements grounded in complex network theory, providing a novel quantitative approach for investigating the mapping relationship between user cognition and design elements within the field of Kansei engineering. First, products are modeled as nodes, and edges are established based on thresholded attribute similarity between products to construct a product attribute network. Second, community detection algorithms are applied to partition the constructed network, revealing the intrinsic association strength among product attributes and enabling the identification of key products within each cluster. Subsequently, a random forest method is employed to extract the core product attributes that influence cluster formation, yielding attribute importance rankings, which are further validated using the XGBoost algorithm to ensure robustness. Finally, a Bluetooth speaker case study is conducted to verify the effectiveness and feasibility of the proposed method for identifying key product design factors. The results demonstrate that the proposed approach can effectively identify key product design factors and rank the critical factors influencing the product design process, achieving high accuracy and practical applicability. Full article

Background: Q&A corpora generated through university student–AI mental health support tools may reveal how physical activity (PA) and psychosocial themes are represented in support-oriented text. However, the absence of individual-level demographic metadata and the pooling of prompt and response fields limit attribution of any expression to a particular speaker, and the corpus describes a specific student population rather than a general or clinical one. Objective: This exploratory study described PA-, sport-, physical education (PE)-, body-, lifestyle-, and emotion-related patterns in a large corpus of university student–AI mental health exchanges collected through an institutional counselling platform. Methods: This study analysed 209,715 paired prompt–response records as combined exchange-level units using a BERTopic-based computational text-mining workflow. The full corpus was used for the main 18-topic model and overlapping dictionary analyses. After secondary data-quality filtering, 178,062 eligible exchanges formed the sampling frame from which a systematic sample of 10,000 exchanges was drawn for a separate complementary BERTopic and scenario-mapping analysis. The workflow used Qdrant/bge-small-zh-v1.5 embeddings, NFKC normalisation, an archived stop-word list, UMAP (n_neighbors = 15, n_components = 5, min_dist = 0.0, cosine metric, seed = 42), HDBSCAN (min_cluster_size = 300, min_samples = 10, Euclidean metric, EOM), c-TF-IDF topic representations, overlapping dictionary screens, and stability testing across seeds 42, 52, and 62. Results: A student/school/family-context lexical screen matched 83,215 exchanges (39.68%), and a broad PA/body/lifestyle screen matched 82,464 exchanges (39.32%). These overlapping indicators describe topical co-occurrence and do not establish PA behaviour or which party to the exchange produced a given term. Eighteen corpus-level themes were retained. In the 10,000-exchange analysis, 13.11% of exchanges matched a narrow movement-related expression screen, with the highest within-topic rate in the sample topic labelled emotional outburst and relaxation regulation (51.09%). Conclusions: The findings describe exchange-level lexical and topic patterns in student–AI interactions rather than actual PA behaviour, intervention delivery, clinical efficacy, or population prevalence, and they do not identify which party introduced the language. The mapping to autonomy, competence, relatedness, and emotional regulation is a post hoc interpretive lens, offered as a hypothesis to inform future, prospectively validated design work in PE and digital mental health support rather than as a demonstrated result. Full article

Audience reactions to oral communication are shaped by both communicative practices and broader social contexts. While elements such as message content, delivery style, and vocal expression can be developed through training, other factors—such as gender and ethnicity—reflect social identities that are often associated with how speakers are perceived and evaluated. This study examines how these contextual attributes are associated with audience engagement in digital public speaking environments. Drawing on an initial dataset of 977 TEDx talks, resulting in two high-confidence subsamples of 610 speakers for gender and 387 for ethnicity, curated through a combination of computational methods with a communication perspective. We analyzed the relationship between the two factors with engagement indicators—including likes, dislikes and interaction rates. The analysis explores whether patterns of audience response differ across demographic groups and at the intersection of gender and ethnicity. The findings reveal that neither gender nor ethnicity, considered on its own, was significantly associated with audience engagement; differences emerged only at the intersection of the two. Specifically, non-Hispanic Black speakers were associated with higher levels of negative feedback in both genders, Hispanic male speakers received more positive engagement than other male speakers, and Asian female speakers showed lower interaction levels—fewer views, likes, and comments—than non-Hispanic White female speakers. These patterns suggest that disparities in how audiences respond to speakers’ social identities in mediated contexts are intersectional, becoming visible only when gender and ethnicity are considered jointly. By providing empirical evidence from a diverse digital corpus, this study contributes to ongoing debates on digital inequalities, representation, and participation in contemporary media environments, highlighting the importance of considering social context in analyses of audience behavior. Full article

Modern Greek was spoken along the northwestern coast of Asia Minor until the early 20th century, yet neither its precise geographical extent nor its dialectal classification is well established. This paper seeks to clarify both issues by focusing on the variety of Adramytti (Edremit). The available evidence suggests that Adramyttian, despite its close relationship to and partial origin in Modern Lesbian, was essentially a mixed variety that leveled out many characteristic Modern Lesbian features, such as the raising of unstressed mid vowels and certain morphological phenomena. Such differences can be attributed to the diverse character of the speech community that led to contact between speakers of Modern Lesbian origin and speakers of other Greek dialects. In addition to providing a grammatical description of Adramyttian, which demonstrates its mixed profile, the paper offers a tentative classification of this variety in relation to Modern Lesbian and the other insular varieties of northeastern Aegean, as well as in relation to other neighboring varieties of northwestern Asia Minor (Aeolis, Mysia, northern Ionia). Full article

In multi-turn conversational AI, individually innocuous personally identifiable information (PII) fragments disclosed across successive turns can accumulate into a re-identification risk that no single utterance reveals on its own. Existing PII detectors operate on isolated utterances and therefore cannot track this cross-turn evidence build-up. We propose a stateful middleware guardrail whose core design principle is speaker-attributed entity isolation: every extracted PII fragment is attributed to its originating conversational participant, and evidence is accumulated in entity-isolated subgraphs that prevent cross-entity contamination. The system signals re-identification onset $t_{\mathrm{pred}}$ at the earliest turn where combination-based rules grounded in the uniqueness literature are satisfied. On a 184-record template-synthetic evaluation corpus, the gated NER configuration leads on primary timeliness (OW@5 $= 73.4\%$, MAE$= 1.357$ turns); the full system achieves OW@5 $= 70.7\%$ with MAE $= 2.442$ turns as an alternative operating mode for ambiguity-sensitive disclosure patterns. We further evaluate behavior on a 300-record mutation stress set, test RULE_B on the ABCD external corpus, and supplement RULE_A evaluation with both a proxy-labeled transfer analysis on PersonaChat and a manual annotation study on 151 Switchboard dialogues. The reported results should be interpreted as an initial empirical reference point rather than a sufficient endpoint for autonomous runtime enforcement. Full article

The detection of deepfake speech has become a significant research area due to rapid advancements in generative AI for speech synthesis. These technologies pose significant security risks in applications such as biometric authentication, voice-controlled systems, and automatic speaker verification (ASV) systems. Therefore, enhancing the detection capabilities of such applications is essential to mitigate potential threats. This study investigates perceptual speech-pathological features, which are commonly used to evaluate the unnaturalness of voice disorders in clinical settings, as potential indicators for detecting deepfake speech. Specifically, the timbral attributes of hardness, depth, brightness, roughness, sharpness, warmth, boominess, and reverberation are examined. The analysis reveals that these attributes provide meaningful distinctions between genuine and synthetic speech. Furthermore, the detection performance is enhanced by extending the dimensional representation of timbral attributes, enabling a more comprehensive characterization of the speech signal. This paper proposes a method that combines two models: one utilizing the different dimensions of speech-pathological features with a deep neural network (DNN), and another employing a gammatone filterbank model that simulates the auditory processing mechanism of the human cochlea with ResNet-18 architecture, improving deepfake speech detection. The proposed method is evaluated on the Automatic Speaker Verification Spoofing and Countermeasures Challenge (ASVspoof) 2019 dataset. Experimental results demonstrate that the proposed approach outperforms baseline models in terms of Equal Error Rate (EER), achieving an EER of $5.93$%. Full article

Despite having approximately 24 million native speakers, Romanian remains a low-resource language for automatic speech recognition (ASR), with few accurate and publicly available systems. To address this gap, this study explores the challenges of adapting modern speech recognition models, such as wav2vec 2.0 and Conformer, to Romanian. Our investigation is a comprehensive analysis of the two models, their capabilities to adapt to Romanian data, and the performance of the trained models. The research also focuses on unique attributes of the Romanian language, data collection techniques, including weakly supervised learning, and processing methodologies. Building on the previously introduced Echo dataset of 378 h, we release CRoWL (Crawled Romanian Weakly Labeled), a weakly supervised dataset of 9000 h created via automatic transcription. We obtain strong results that, to the best of our knowledge, are competitive with or exceed publicly reported results for Romanian under comparable open evaluation settings, with Conformer attaining 3.01% WER on Echo + CRoWL and wav2vec 2.0 reaching 4.04% (Echo) and 4.17% (Echo + CRoWL). In addition to the datasets, we also release our most capable models as open source, along with their training plans, thereby providing a solid foundation for researchers interested in languages with limited representation. Full article

As generative speech synthesis produces near-human synthetic voices and reliance on online media grows, robust audio-deepfake detection is essential to fight misuse and misinformation. In this study, we introduce the Arabic Fake Audio Dataset for Modern Standard Arabic (AFAD-MSA), a curated corpus of authentic and synthetic Arabic speech designed to advance research on Arabic deepfake and spoofed-speech detection. The synthetic subset is generated with four state-of-the-art proprietary text-to-speech and voice-conversion models. Rich metadata—covering speaker attributes and generation information—is provided to support reproducibility and benchmarking. To establish reference performance, we trained three AASIST models and compared their performance to two baseline transformer detectors (Wav2Vec 2.0 and Whisper). On the AFAD-MSA test split, AASIST-2 achieved perfect accuracy, surpassing the baseline models. However, its performance declined under cross-dataset evaluation. These results underscore the importance of data construction. Detectors generalize best when exposed to diverse attack types. In addition, continual or contrastive training that interleaves bona fide speech with large, heterogeneous spoofed corpora will further improve detectors’ robustness. Full article

This paper presents Fic2Bot, an end-to-end framework that automatically transforms raw novel text into in-character chatbots by combining scene-level retrieval with persona profiling. Unlike conventional RAG-based systems that emphasize factual accuracy but neglect stylistic coherence, Fic2Bot ensures both factual grounding and consistent persona expression without any manual intervention. The framework integrates (1) Major Entity Identification (MEI) for robust coreference resolution, (2) scene-structured retrieval for precise contextual grounding, and (3) stylistic and sentiment profiling to capture linguistic and emotional traits of each character. Experiments conducted on novels from diverse genres show that Fic2Bot achieves robust entity resolution, more relevant retrieval, highly accurate speaker attribution, and stronger persona consistency in multi-turn dialogues. These results highlight Fic2Bot as a scalable and domain-agnostic framework for persona-driven chatbot generation, with potential applications in interactive roleplaying, language and literary studies, and entertainment. Full article

Language ideologies about gendered linguistic behaviour are crucial in shaping expectations and metapragmatic judgements on politeness. This study focused on how gender and language ideologies reinforce normative assumptions about the relationship between gender and (im)politeness and at the same time influence individuals’ perception of (im)politeness. Based on data collected from 251 respondents through online questionnaires administered between July 2024 and January 2025, the study investigated whether certain linguistic choices tend to be stereotypically associated with a particular gender and if the same utterance is evaluated differently depending on whether it is attributed to a man or a woman. Participants’ responses revealed systematic associations between linguistic forms and perceived gender, indicating that direct requests were more often linked to male speakers, while indirect or mitigated forms were associated with female speakers. Findings also showed that in 17 out of 19 cases, the same utterance was rated as more polite when attributed to a woman, suggesting that among Russian-speaking participants politeness was not only expected from women but also more readily perceived in their speech, reinforcing existing gender ideologies and stereotypes. Full article

The aim of this study is to provide a systematic account of structural discourse markers operating at a conversational macro-level in German Palliative Care interactions, focusing on their frequency, distribution, co-occurrence, and speaker-group-specific usage. By combining qualitative approaches from conversation analysis and interactional linguistics with quantitative methods from corpus linguistics, discourse markers are analyzed together as a functional category from multiple analytical perspectives to enhance the overall understanding of the use of discourse markers. The analysis reveals a functional distribution across different transition points in conversation: Whereas the German so most frequently appears in openings and transitions to non-verbal activities, gut and okay predominate in topic shifts and conversation closings. However, gut and okay differ in their composition of discursive functions, although discourse structuring emerges as the second most frequent function in both cases, an observation that stands in contrast to the continued neglect of this function in standard dictionary entries. The institutional asymmetries inherent in medical interactions are reflected in the finding that both doctors and caregivers use significantly more structural discourse markers than patients and their relatives. Differences between physicians’ and caregivers’ use of discourse markers can be attributed to their different professional roles and communicative responsibilities. Full article

Speaker burn-in is a controversial practice in the audio world, based on the belief that new devices reach optimal performance only after a certain period of use. Supporters claim it improves component flexibility, reduces initial distortion, and enhances sound quality—especially in the low-frequency range. Critics, however, emphasize the lack of scientific evidence for audible changes and point to the placebo effect in subjective listening tests. They argue that modern manufacturing and strict quality control minimize differences between new and “burned-in” devices. This study cites a standard describing a preliminary burn-in procedure, specifying the exact conditions and duration required. Objective tests revealed slight changes in speaker impedance and amplitude response after burn-in, but these differences are inaudible to the average listener. Notably, significant variation was observed between speakers of the same series, attributed to production line tolerances rather than use-related changes. The study also explored aging processes in speaker materials to better understand potential long-term effects. However, subjective listening tests showed that listeners rated the sound consistently across all test cases, regardless of whether the speaker had undergone burn-in. Overall, while minor physical changes may occur, their audible impact is negligible, especially for non-expert users. Full article

The present study investigates how US Southerners perceive second language (L2) speech by recruiting 170 undergraduate students who spoke Southern American English to listen to recordings of four speakers (US, Bangladeshi, Chinese, and Saudi Arabian) and evaluate their attributes. The listeners were grouped based on their ethnic affiliation: African American, Anglo-American, and Asian/Hispanic/multi-racial. A random half were primed, being asked questions about whether/how other people had negatively commented on their accents. Results showed no effect of priming on speech ratings. Moreover, whilst African American and Anglo-American listeners rated L2 speakers lower than the L1 speaker in almost all aspects, Asian/Hispanic/multi-racial listeners did not. Full article

Objectives: We introduce the conceptual framework for the Adversarial Claim Robustness Diagnostics (ACRD) protocol, a novel tool for assessing how factual claims withstand ideological distortion. Methods: Based on semantics, adversarial collaboration, and the devil’s advocate approach, we develop a three-phase evaluation process combining baseline evaluations, adversarial speaker reframing, and dynamic AI calibration along with quantified robustness scoring. We introduce the Claim Robustness Index that constitutes our final validity scoring measure. Results: We model the evaluation of claims by ideologically opposed groups as a strategic game with a Bayesian-Nash equilibrium to infer the normative behavior of evaluators after the reframing phase. The ACRD addresses shortcomings in traditional fact-checking approaches and employs large language models to simulate counterfactual attributions while mitigating potential biases. Conclusions: The framework’s ability to identify boundary conditions of persuasive validity across polarized groups can be tested across important societal and political debates ranging from climate change issues to trade policy discourses. Full article

Studies that have examined /s/ weakening as a social practice have found that L1 Spanish speakers perceive this cue as an indicator of lower status, region of origin, and greater friendliness, and even L2 Spanish learners have been found to associate /s/ weakening with lower status. The question remains, however, whether L2 learners who use /s/ weakening are perceived as having these same social attributes or whether their nonnative status interrupts said assessment. The present study examines the attitudes of 30 beginning and intermediate-level L2 learners of Spanish towards L1 and L2 speech that was digitally modified to contain /s/ deletion in coda positions, a regionally and stylistically variable sociophonetic cue in Spanish that is often not adopted by L2 learners. Participants rated the speech samples on Likert scales of perceived status, solidarity, and nativeness. Results revealed that learners rated L1 speech with /s/ deletion significantly lower for status and L2 speech with /s/ deletion significantly higher for nativeness. Full article