A Hybrid Usability Approach for Rating Evaluation of M-Commerce Applications
研究通过分析八个可用性因素及用户数据，提出一种基于逐步多元线性回归的混合可用性模型，以预测M-Commerce应用的评分。
0 citationsRead paper
Selected work
Representative Papers
A Controlled Study of CLIP-Based Body-Scene Fusion for Emotion Recognition in Context
This work addresses emotion recognition in natural images by jointly modeling human body posture and scene context. The authors propose a dual-stream architecture that encodes the body region using ResNet-18 and the entire scene using CLIP’s ViT-B/16, followed by fusion to predict 26 discrete emotion categories along with continuous valence, arousal, and dominance dimensions. Comprehensive evaluation reveals that, after incorporating CLIP-based scene representations, various debiasing and few-shot strategies—including causal intervention, CLEF-lite, and ASL tuning—fail to improve performance. The base dual-stream model achieves 34.52% mAP on the EMOTIC test set, outperforming all variants, suggesting that CLIP already captures sufficient scene semantics. Future efforts should therefore focus on modeling label relationships and fine-grained subject-scene interactions.
0 citationsRead paper
Irony Detection in Urdu Text: A Comparative Study Using Machine Learning Models and Large Language Models
This work addresses the low-resource challenge of Urdu sarcasm detection by proposing a cross-lingual transfer framework for data construction and model evaluation. We translate high-quality English sarcasm corpora into Urdu and perform rigorous human verification, thereby establishing the first publicly available, consistently annotated Urdu sarcasm dataset. Methodologically, we systematically compare traditional word embeddings (GloVe, Word2Vec), multilingual pre-trained language models (mBERT, XLM-RoBERTa), and leading open-source large language models (LLaMA 2/3, Mistral), employing both fine-tuning and embedding-based strategies. Crucially, we integrate Urdu transliteration techniques with modern NLP architectures to bridge a critical gap in Urdu sarcasm recognition research. Experimental results demonstrate that fine-tuned LLaMA-3 (8B) achieves an F1-score of 94.61%, significantly outperforming the gradient-boosting baseline (89.18%), thus validating the efficacy of large language models for low-resource sarcasm detection.
0 citationsRead paper
Do LLMs Know They Are Being Tested? Evaluation Awareness and Incentive-Sensitive Failures in GPT-OSS-20B
This paper identifies a pervasive “evaluation bias” in LLM benchmarking: current prompts frequently embed explicit scoring cues—such as enforced chain-of-thought reasoning or rigid formatting constraints—distorting model behavior and inflating performance metrics without genuine capability improvement. To systematically investigate this, the authors design a reproducible A/B testing framework and conduct six controlled experiments on the GPT-OSS-20B model. Using deterministic validators, structured parsing, and multidimensional evaluation metrics, they analyze the impact of task framing and reasoning depth. Results show that evaluation-aware prompting induces redundant reasoning chains and reduced compliance; subtle changes in incentive wording shift error type distributions; and non-English prompts significantly degrade performance on high-reasoning tasks. Crucially, the study provides the first empirical evidence that evaluation-oriented prompting fails to enhance accuracy stability and instead introduces systematic biases.
0 citationsRead paper
Cuffless Blood Pressure Prediction from Speech Sentences using Deep Learning Methods
Conventional cuff-based blood pressure (BP) monitoring suffers from white-coat and masked hypertension effects, poor patient compliance, and limited suitability for continuous or remote assessment. To address these limitations, this study proposes a non-invasive, cuffless BP estimation method leveraging speech signals. We introduce the first application of the BERT architecture to BP regression from speech, integrating acoustic feature extraction with deep learning in an end-to-end framework. Evaluated on speech recordings from 95 subjects, the model achieves mean absolute errors of 13.6 mmHg for systolic BP and 12.4 mmHg for diastolic BP, with R² scores of 0.99 and 0.94, respectively. The approach demonstrates strong generalizability and training stability across diverse participants. By circumventing reliance on traditional physiological signals (e.g., ECG, PPG), this method establishes a scalable, user-friendly paradigm for telemedicine and long-term cardiovascular health monitoring.
0 citationsRead paper
Recent publications
Latest Papers
A Hybrid Usability Approach for Rating Evaluation of M-Commerce Applications
研究通过分析八个可用性因素及用户数据，提出一种基于逐步多元线性回归的混合可用性模型，以预测M-Commerce应用的评分。
0 citationsRead paper
A Controlled Study of CLIP-Based Body-Scene Fusion for Emotion Recognition in Context
This work addresses emotion recognition in natural images by jointly modeling human body posture and scene context. The authors propose a dual-stream architecture that encodes the body region using ResNet-18 and the entire scene using CLIP’s ViT-B/16, followed by fusion to predict 26 discrete emotion categories along with continuous valence, arousal, and dominance dimensions. Comprehensive evaluation reveals that, after incorporating CLIP-based scene representations, various debiasing and few-shot strategies—including causal intervention, CLEF-lite, and ASL tuning—fail to improve performance. The base dual-stream model achieves 34.52% mAP on the EMOTIC test set, outperforming all variants, suggesting that CLIP already captures sufficient scene semantics. Future efforts should therefore focus on modeling label relationships and fine-grained subject-scene interactions.
0 citationsRead paper
Irony Detection in Urdu Text: A Comparative Study Using Machine Learning Models and Large Language Models
This work addresses the low-resource challenge of Urdu sarcasm detection by proposing a cross-lingual transfer framework for data construction and model evaluation. We translate high-quality English sarcasm corpora into Urdu and perform rigorous human verification, thereby establishing the first publicly available, consistently annotated Urdu sarcasm dataset. Methodologically, we systematically compare traditional word embeddings (GloVe, Word2Vec), multilingual pre-trained language models (mBERT, XLM-RoBERTa), and leading open-source large language models (LLaMA 2/3, Mistral), employing both fine-tuning and embedding-based strategies. Crucially, we integrate Urdu transliteration techniques with modern NLP architectures to bridge a critical gap in Urdu sarcasm recognition research. Experimental results demonstrate that fine-tuned LLaMA-3 (8B) achieves an F1-score of 94.61%, significantly outperforming the gradient-boosting baseline (89.18%), thus validating the efficacy of large language models for low-resource sarcasm detection.
0 citationsRead paper
Do LLMs Know They Are Being Tested? Evaluation Awareness and Incentive-Sensitive Failures in GPT-OSS-20B
This paper identifies a pervasive “evaluation bias” in LLM benchmarking: current prompts frequently embed explicit scoring cues—such as enforced chain-of-thought reasoning or rigid formatting constraints—distorting model behavior and inflating performance metrics without genuine capability improvement. To systematically investigate this, the authors design a reproducible A/B testing framework and conduct six controlled experiments on the GPT-OSS-20B model. Using deterministic validators, structured parsing, and multidimensional evaluation metrics, they analyze the impact of task framing and reasoning depth. Results show that evaluation-aware prompting induces redundant reasoning chains and reduced compliance; subtle changes in incentive wording shift error type distributions; and non-English prompts significantly degrade performance on high-reasoning tasks. Crucially, the study provides the first empirical evidence that evaluation-oriented prompting fails to enhance accuracy stability and instead introduces systematic biases.
0 citationsRead paper
Cuffless Blood Pressure Prediction from Speech Sentences using Deep Learning Methods
Conventional cuff-based blood pressure (BP) monitoring suffers from white-coat and masked hypertension effects, poor patient compliance, and limited suitability for continuous or remote assessment. To address these limitations, this study proposes a non-invasive, cuffless BP estimation method leveraging speech signals. We introduce the first application of the BERT architecture to BP regression from speech, integrating acoustic feature extraction with deep learning in an end-to-end framework. Evaluated on speech recordings from 95 subjects, the model achieves mean absolute errors of 13.6 mmHg for systolic BP and 12.4 mmHg for diastolic BP, with R² scores of 0.99 and 0.94, respectively. The approach demonstrates strong generalizability and training stability across diverse participants. By circumventing reliance on traditional physiological signals (e.g., ECG, PPG), this method establishes a scalable, user-friendly paradigm for telemedicine and long-term cardiovascular health monitoring.
0 citationsRead paper