An official website of the United States government
The use of machine learning techniques has significantly increased the physics discovery potential of neutrino telescopes. In the upcoming years, we are expecting upgrades of currently existing detectors and new telescopes with novel experimental hardware, yielding more statistics as well as more complicated data signals. This calls for an upgrade on the software side needed to handle this more complex data in a more efficient way. Specifically, we seek low power and fast software methods to achieve real-time signal processing, where current machine learning methods are too expensive to be deployed in the resource-constrained regions where these experiments are located. We present the first attempt at and a proof-of-concept for enabling machine learning methods to be deployed in-detector for water/ice neutrino telescopes via quantization and deployment on Google Edge Tensor Processing Units (TPUs). We design a recursive neural network with a residual convolutional embedding and adapt a quantization process to deploy the algorithm on a Google Edge TPU. This algorithm can achieve similar reconstruction accuracy compared with traditional GPU-based machine learning solutions while requiring the same amount of power compared with CPU-based regression solutions, combining the high accuracy and low power advantages and enabling real-time in-detector machine learning in even the most power-restricted environments.
more »
« less
This content will become publicly available on March 22, 2026
Real-Time Posture Classification on Resource-Constrained Runtime Environments: A Comparative Study of Traditional and Deep Learning Models
While machine learning models perform well on offline data, assessing their performance in real-world, resource-constrained environments-considering accuracy, prediction time, power consumption, and memory usage-is crucial for practical applications. This research implements a mobile-based Human Activity Recognition solution to classify three postures-sitting, standing, and walking-using smartphone sensors, specifically accelerometer, gyroscope, and magnetometer. Time-domain features extracted from these sensors were used, with Random Forest employed for feature selection. One traditional machine learning model, Logistic Regression, and one deep learning model, Convolutional Neural Network, were trained and deployed via an Android application for real-time evaluation. While the Convolutional Neural Network achieved higher accuracy and better memory efficiency, Logistic Regression demonstrated faster prediction times during real-time use. Both models showed reduced accuracy for standing and walking postures in real-world conditions, emphasizing the challenges of deploying machine learning models in dynamic environments. This study highlights the importance of evaluating machine learning models in real-world settings to ensure reliability and efficiency, particularly in resource-constrained environments.
more »
« less
- Award ID(s):
- 2131100
- PAR ID:
- 10585203
- Publisher / Repository:
- IEEE
- Date Published:
- ISBN:
- 979-8-3315-0484-7
- Page Range / eLocation ID:
- 1162 to 1170
- Format(s):
- Medium: X
- Location:
- Concord, NC, USA
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Abstract Milling is a critical manufacturing process to produce high-value components in aerospace, tooling, and automotive industries. However, milling is prone to chatter, a severe vibration that damages surface quality, cutting tools, and machines. Traditional experimental and mechanistic methods of chatter prediction have significant limitations. This study presents a data-driven machine learning (ML) model to predict and quantify milling chatter directly based on time-series vibration data. Three ML models, including hybrid long short-term memory (LSTM)—fully convolutional network (FCN) model, gated recurrent unit (GRU)—FCN model, and temporal convolutional network (TCN) models, have been developed and verified by incorporating milling parameters to enhance prediction accuracy and stability. Among the proposed models, the best-performing ML model (GRU-FCN) demonstrates strong performance in chatter prediction and severity quantification, providing actionable insights with improved computational efficiency. The integration of milling parameters into the ML model notably enhances the prediction accuracy and stability, proving particularly effective in real-time monitoring scenarios.more » « less
-
With the increasing use of Unmanned Aerial Vehicles in military and civilian applications, the security of this technology has become one of the critical concerns. UAVs’ positioning and navigation activities are highly dependent on Global Positioning Systems as they provide accurate locations for these vehicles. However, due to the civilian GPS signals being open and unencrypted, malicious users can target them in multiple ways, including by launching Global Positioning System spoofing attacks. To address this security issue, numerous techniques have been proposed to detect and classify these attacks, including supervised machine learning techniques. However, no studies have focused on unsupervised models to detect these attacks. In this paper, we compare the performance of several supervised models with that of unsupervised models in terms of accuracy, probability of detection, probability of misdetection, probability of false alarm, processing time, training time, prediction time, and memory size. The supervised models are Gaussian Naïve Bayes, Classification and Regression Decision Tree, Logistic Regression, Random Forest, Linear-Support Vector Machine, and Artificial Neural Network. The unsupervised models are Principal Component Analysis, K-means clustering, and Autoencoder. The results show that the Classification and Regression Decision Tree model outperforms the other supervised and unsupervised models in detecting and classifying GPS spoofing attacks.more » « less
-
null (Ed.)Background : Machine learning has been used for classification of physical behavior bouts from hip-worn accelerometers; however, this research has been limited due to the challenges of directly observing and coding human behavior “in the wild.” Deep learning algorithms, such as convolutional neural networks (CNNs), may offer better representation of data than other machine learning algorithms without the need for engineered features and may be better suited to dealing with free-living data. The purpose of this study was to develop a modeling pipeline for evaluation of a CNN model on a free-living data set and compare CNN inputs and results with the commonly used machine learning random forest and logistic regression algorithms. Method : Twenty-eight free-living women wore an ActiGraph GT3X+ accelerometer on their right hip for 7 days. A concurrently worn thigh-mounted activPAL device captured ground truth activity labels. The authors evaluated logistic regression, random forest, and CNN models for classifying sitting, standing, and stepping bouts. The authors also assessed the benefit of performing feature engineering for this task. Results : The CNN classifier performed best (average balanced accuracy for bout classification of sitting, standing, and stepping was 84%) compared with the other methods (56% for logistic regression and 76% for random forest), even without performing any feature engineering. Conclusion : Using the recent advancements in deep neural networks, the authors showed that a CNN model can outperform other methods even without feature engineering. This has important implications for both the model’s ability to deal with the complexity of free-living data and its potential transferability to new populations.more » « less
-
Embedded devices, constrained by limited memory and processors, require deep learning models to be tailored to their specifications. This research explores customized model architectures for classifying drainage crossing images. Building on the foundational ResNet-18, this paper aims to maximize prediction accuracy, reduce memory size, and minimize inference latency. Various configurations were systematically probed by leveraging hardware-aware neural architecture search, accumulating 1,717 experimental results over six benchmarking variants. The experimental data analysis, enhanced by nn-Meter, provided a comprehensive understanding of inference latency across four different predictors. Significantly, a Pareto front analysis with three objectives of accuracy, latency, and memory resulted in five non-dominated solutions. These standout models showcased efficiency while retaining accuracy, offering a compelling alternative to the conventional ResNet-18 when deployed in resource-constrained environments. The paper concludes by highlighting insights drawn from the results and suggesting avenues for future exploration.more » « less
