Selected work

Representative Papers

SMART: A Machine Learning and Monte Carlo Framework for Rapid Analysis of Stochastic Transistor Aging and Process Variation in Digital Circuits

Jul 06, 2026

This work addresses the significant challenge posed by bias temperature instability (BTI) and random process–voltage (PV) variations to digital circuit reliability in deep-nanometer CMOS technologies, where conventional analysis methods suffer from high computational cost and poor scalability. To overcome these limitations, the paper proposes a novel gate-level delay distribution prediction framework that uniquely integrates random forest regression with Bayesian optimization. By leveraging offline training on Monte Carlo simulation data—bypassing time-consuming atomic parameter extraction—and employing Bayesian optimization for automated hyperparameter tuning, the approach achieves substantially improved accuracy and efficiency. Experimental validation on ISCAS85 benchmark circuits demonstrates a 94.54% reduction in analysis time compared to the state-of-the-art method, with an average prediction error of only 1.63%.

0 citationsRead paper

EEG Sleep Stage Classification with Continuous Wavelet Transform and Deep Learning

Oct 08, 2025

To address the time-consuming nature of manual sleep-stage annotation and the poor interpretability of traditional handcrafted-feature methods, this paper proposes an automated sleep-stage classification framework integrating continuous wavelet transform (CWT) with deep ensemble learning. CWT is employed to generate high-resolution time-frequency representations that faithfully capture stage-specific transient activities and rhythmic oscillations. A convolutional neural network (CNN) is then applied to extract discriminative local time-frequency patterns, while an ensemble strategy enhances model robustness and decision interpretability. Evaluated on the Sleep-EDF dataset, the framework achieves an overall accuracy of 88.37% and a macro-averaged F1-score of 73.15%, matching state-of-the-art deep learning methods and significantly outperforming conventional machine learning approaches. Crucially, the method preserves clinical interpretability through physiologically grounded time-frequency features, offering both high predictive performance and transparency for practical deployment in sleep medicine.

0 citationsRead paper

Recent publications

Latest Papers

SMART: A Machine Learning and Monte Carlo Framework for Rapid Analysis of Stochastic Transistor Aging and Process Variation in Digital Circuits

Jul 06, 2026

This work addresses the significant challenge posed by bias temperature instability (BTI) and random process–voltage (PV) variations to digital circuit reliability in deep-nanometer CMOS technologies, where conventional analysis methods suffer from high computational cost and poor scalability. To overcome these limitations, the paper proposes a novel gate-level delay distribution prediction framework that uniquely integrates random forest regression with Bayesian optimization. By leveraging offline training on Monte Carlo simulation data—bypassing time-consuming atomic parameter extraction—and employing Bayesian optimization for automated hyperparameter tuning, the approach achieves substantially improved accuracy and efficiency. Experimental validation on ISCAS85 benchmark circuits demonstrates a 94.54% reduction in analysis time compared to the state-of-the-art method, with an average prediction error of only 1.63%.

0 citationsRead paper

EEG Sleep Stage Classification with Continuous Wavelet Transform and Deep Learning

Oct 08, 2025

To address the time-consuming nature of manual sleep-stage annotation and the poor interpretability of traditional handcrafted-feature methods, this paper proposes an automated sleep-stage classification framework integrating continuous wavelet transform (CWT) with deep ensemble learning. CWT is employed to generate high-resolution time-frequency representations that faithfully capture stage-specific transient activities and rhythmic oscillations. A convolutional neural network (CNN) is then applied to extract discriminative local time-frequency patterns, while an ensemble strategy enhances model robustness and decision interpretability. Evaluated on the Sleep-EDF dataset, the framework achieves an overall accuracy of 88.37% and a macro-averaged F1-score of 73.15%, matching state-of-the-art deep learning methods and significantly outperforming conventional machine learning approaches. Crucially, the method preserves clinical interpretability through physiologically grounded time-frequency features, offering both high predictive performance and transparency for practical deployment in sleep medicine.

0 citationsRead paper