Transport Mode Detection (TMD) systems play a pivotal role in facilitating applications in transport, urban planning, and more. Exploiting the advancements in smartphone sensing capabilities, TMD systems have evolved ...
详细信息
ISBN:
(纸本)9798331528690;9798331528706
Transport Mode Detection (TMD) systems play a pivotal role in facilitating applications in transport, urban planning, and more. Exploiting the advancements in smartphone sensing capabilities, TMD systems have evolved for mobile applications with local classification on smartphones as a common approach. Yet, local approaches relying on centralized training raise privacy concerns due to the transmission of sensitive data (e.g., GPS logs) over the Internet. In this paper, we propose FOGFLEET, a novel Federated Transfer learning (FTL) framework for TMD, addressing both privacy and performance concerns. Our approach relies on Federated learning (FL) to train a global model on various datasets from different cities while employing transfer learning to adapt the global model to the specific characteristics of individual smartphones and cities. FOGFLEET relies on an architecture that integrates edge, fog, and cloud layers, with dedicated fog nodes for each city to simplify cross-silo federated learning. Experimental results demonstrate the effectiveness of the FOGFLEET framework in higher TMD accuracy by up to 20% than its comparable centralized approach. Furthermore, it outperforms the FL solutions reported in the literature with at least an 8% increase in accuracy. In this work, we also highlight the importance of sufficient training data for distributed training and discuss the impact of smartphone sensor qualities on the performance of TMD systems. Our work contributes to advancing TMD systems by providing an adaptive and privacy-preserving solution suitable for deployment in diverse urban environments and across various geographical locations.
This paper describes the components and configurations available in a new quantum-assisted machine learning (QAML) framework. QAML is an open source package that provides a new and flexible test-bed for algorithms to ...
详细信息
ISBN:
(纸本)9798331541378
This paper describes the components and configurations available in a new quantum-assisted machine learning (QAML) framework. QAML is an open source package that provides a new and flexible test-bed for algorithms to train and evaluate Boltzmann machines (BMs) using quantum annealers. Quantum annealing processors enable the training capabilities for both restricted (RBM) and general Boltzmann machines (BM). These methods rely on the fidelity of samples from those devices, which approximate the characteristic Boltzmann distribution of the BM models. The models and optimization functions are built on top of the PyTorch and D-Wave Ocean libraries. The goal of this paper is to introduce developers and researchers to a familiar, yet flexible, software development framework, to explore the use of quantum computing in machine learning applications. The framework is open-source and has been made publicly available.
Breakthroughs in machine learning (ML) and advances in quantum computing (QC) drive the interdisciplinary field of quantum machine learning to new levels. However, due to the susceptibility of ML models to adversarial...
详细信息
ISBN:
(纸本)9798331541378
Breakthroughs in machine learning (ML) and advances in quantum computing (QC) drive the interdisciplinary field of quantum machine learning to new levels. However, due to the susceptibility of ML models to adversarial attacks, practical use raises safety-critical concerns. Existing Randomized Smoothing (RS) certification methods for classical machine learning models are computationally intensive. In this paper, we propose the combination of QC and the concept of discrete randomized smoothing to speed up the stochastic certification of ML models for discrete data. We show how to encode all the perturbations of the input binary data in superposition and use Quantum Amplitude Estimation (QAE) to obtain a quadratic reduction in the number of calls to the model that are required compared to traditional randomized smoothing techniques. In addition, we propose a new binary threat model to allow for an extensive evaluation of our approach on images, graphs, and text.
Aiming at the problems of large consumption of evaluation time and low completeness of evaluation information in traditional evaluation methods, this paper puts forward a satisfaction evaluation method of compound tea...
详细信息
Aiming at the problems of large consumption of evaluation time and low completeness of evaluation information in traditional evaluation methods, this paper puts forward a satisfaction evaluation method of compound teaching mode from the perspective of MOOC concept. Firstly, the MOOC curriculum model is analysed, the satisfaction evaluation index system is established for the compound teaching model, and five first-class indexes and 18 second-class indexes are obtained. Then, the evaluation matrix is established to sort the evaluation indexes of teaching satisfaction, calculate the consistency ratio of the index system, and obtain the weight of each evaluation index by means of objective assignment. Finally, cloud computing technology is used to complete the analysis of satisfaction evaluation data. The experimental results show that the time consumption of this method is between 27.62 s and 34.59 s, the completeness of evaluation information is between 0.923 and 0.951, and the accuracy of evaluation results is between 91.7% and 96.3%.
In the current situation of intense global competition and rapid adaptation to new technological, economic and social challenges, junior engineers must be able to contribute to solving business problems in a relativel...
详细信息
ISBN:
(纸本)9783031653995;9783031654008
In the current situation of intense global competition and rapid adaptation to new technological, economic and social challenges, junior engineers must be able to contribute to solving business problems in a relatively short period of time, and exhibit high capacities for continuous learning and adaptation in professional contexts. The learning Factory is a promising approach to competence development. The Dual learning Factory approach in the Process and Product Innovation engineering curriculum of the IMH presented here allows for a systematic and learning-oriented design of engineering competencies by integrating conceptual knowledge and engineering skills in the context of a company. This article focuses on the classroom training part and the teaching-learning method modifications that have been made to develop engineering competencies in the students. Based on a project-based teaching approach, several STEM projects integrated into classroom teaching were designed and developed. The results are encouraging, although they show that there is a need for students to have opportunities to use scientific and technical skills over a long period of time.
In recent years, with the increasing level of global integration and English internationalization, the demand for English learning has also grown rapidly. However, the current deep learning of English is relatively we...
详细信息
The proceedings contain 94 papers. The topics discussed include: deep learning prediction of exonic sequence;factors influencing users’ perspective on adopting cloud computing framework in higher educational institut...
ISBN:
(纸本)9798350327090
The proceedings contain 94 papers. The topics discussed include: deep learning prediction of exonic sequence;factors influencing users’ perspective on adopting cloud computing framework in higher educational institutions of Bangladesh;toxicity detection in water based on polycyclic aromatic hydrocarbons using machine learning algorithms;forecasting the efficacy of e-waste management system using Fourier series approach towards environmental sustainability;implementation of an efficient IoT enabled automated paralysis healthcare system;integration and validation of a robust process for cloud applications;forecasting teaching and learning outcomes effects of mathematics courses in the mechatronics engineering degree at UTB;and goal-oriented prioritized non-functional testing with stakeholders' priorities.
engineering-based learning could bring integrated STEM issues to everyday life through a design process based on a group problem-solving approach. In this study, high school students participated in a mouse-trap car p...
详细信息
ISBN:
(纸本)9781665453318
engineering-based learning could bring integrated STEM issues to everyday life through a design process based on a group problem-solving approach. In this study, high school students participated in a mouse-trap car project that focused on engineering-based STEM learning. A qualitative approach was used to examine how STEM-related subjects were applied by students while they created design sketches and prototypes. Data were gathered through the documenting of student workbooks, direct observations of 28 students in class, and audio and video recordings of those observations. During the prototype design and development phase, the students focused on scientific concepts related to wheel friction and aerodynamics. By modifying the car's structure, adding rubber bands to the wheels to increase friction, and strengthening the force on the levers and rope reels, the students showed their abilities to improve the design. This study demonstrates how practical STEM learning activities such as engineering challenges help students integrate science and math topics in a systematic problem-solving process.
The concept of Outcome-Based Education (OBE) advocates student-centered, outcome-based, and continuous improvement. Under the situation that the structure of students in higher vocational colleges is complex and their...
详细信息
In response to the demand for real-time performance and control quality in industrial Internet of Things (IoT) environments, this paper proposes an optimization control system based on deep reinforcement learning and ...
详细信息
ISBN:
(纸本)9798350366105;9798350366099
In response to the demand for real-time performance and control quality in industrial Internet of Things (IoT) environments, this paper proposes an optimization control system based on deep reinforcement learning and edge computing. The system leverages cloud-edge collaboration, deploys lightweight policy networks at the edge, predicts system states, and outputs controls at a high frequency, enabling monitoring and optimization of industrial objectives. Additionally, a dynamic resource allocation mechanism is designed to ensure rational scheduling of edge computing resources, achieving global optimization. Results demonstrate that this approach reduces cloud-edge communication latency, accelerates response to abnormal situations, reduces system failure rates, extends average equipment operating time, and saves costs for manual maintenance and replacement. This ensures real-time and stable control.
暂无评论