Machine learning for microcontroller-class hardware: A review

SS Saha, SS Sandha, M Srivastava - IEEE Sensors Journal, 2022 - ieeexplore.ieee.org
The advancements in machine learning (ML) opened a new opportunity to bring intelligence
to the low-end Internet-of-Things (IoT) nodes, such as microcontrollers. Conventional ML …

Difformer: Multi-resolutional differencing transformer with dynamic ranging for time series analysis

B Li, W Cui, L Zhang, C Zhu, W Wang… - … on Pattern Analysis …, 2023 - ieeexplore.ieee.org
Time series analysis is essential to many far-reaching applications of data science and
statistics including economic and financial forecasting, surveillance, and automated …

Primenet: Pre-training for irregular multivariate time series

RR Chowdhury, J Li, X Zhang, D Hong… - Proceedings of the …, 2023 - ojs.aaai.org
Real-world applications often involve irregular time series, for which the time intervals
between successive observations are non-uniform. Irregularity across multiple features in a …

TFSemantic: A Time–Frequency Semantic GAN Framework for Imbalanced Classification Using Radio Signals

P Liao, X Wang, L An, S Mao, T Zhao… - ACM Transactions on …, 2024 - dl.acm.org
Recently, wireless sensing techniques have been widely used for Internet of Things (IoT)
applications. Unlike traditional device-based sensing, wireless sensing is contactless …

Multivariate time series classification based on fusion features

M Du, Y Wei, Y Hu, X Zheng, C Ji - Expert Systems with Applications, 2024 - Elsevier
In various areas of real life, Multivariate Time Series Classification (MTSC) is widely used. It
has been the focus of attention of many researchers, and a number of MTSC methods have …

FreqMAE: Frequency-Aware Masked Autoencoder for Multi-Modal IoT Sensing

D Kara, T Kimura, S Liu, J Li, D Liu, T Wang… - Proceedings of the …, 2024 - dl.acm.org
This paper presents FreqMAE, a novel self-supervised learning framework that synergizes
masked autoencoding (MAE) with physics-informed insights to capture feature patterns in …

On the amplification of security and privacy risks by post-hoc explanations in machine learning models

P Quan, S Chakraborty, JV Jeyakumar… - arXiv preprint arXiv …, 2022 - arxiv.org
A variety of explanation methods have been proposed in recent years to help users gain
insights into the results returned by neural networks, which are otherwise complex and …

[HTML][HTML] CIR-DFENet: Incorporating cross-modal image representation and dual-stream feature enhanced network for activity recognition

Y Zhao, J Shao, X Lin, T Sun, J Li, C Lian, X Lyu… - Expert Systems with …, 2025 - Elsevier
Human activity recognition (HAR) based on wearable sensors has been widely used in
various fields such as health monitoring, healthcare, and fitness due to its portability …

NeurIT: Pushing the Limit of Neural Inertial Tracking for Indoor Robotic IoT

X Zheng, S Ji, Y Pan, K Zhang, C Wu - arXiv preprint arXiv:2404.08939, 2024 - arxiv.org
Inertial tracking is vital for robotic IoT and has gained popularity thanks to the ubiquity of low-
cost Inertial Measurement Units (IMUs) and deep learning-powered tracking algorithms …

Centaur: Robust Multimodal Fusion for Human Activity Recognition

S Xaviar, X Yang, O Ardakanian - IEEE Sensors Journal, 2024 - ieeexplore.ieee.org
The proliferation of Internet of Things (IoT) and mobile devices equipped with
heterogeneous sensors has enabled new applications that rely on the fusion of time series …