Special Topics (Smart Structures Health Monitoring and Data Mining)
- Section 1. Advanced Instrumentation and Intelligent Structural Monitoring
Data acquisition systems and various sensor technologies, remote monitoring methods and non-destructive testing, real-world examples of bridge instrumentation, data-driven structural health monitoring and smart structures, emerging technologies such as the Internet of Things (IoT), blockchain, and digital twins, as well as limitations, risks, lessons learned, and future outlook.
- Section 2. Deep Learning-Based Data Mining Techniques:
A brief overview of some classical machine learning methods, model evaluation, anomaly detection, neural network architectures, multilayer perceptron (MLP), loss functions, forward propagation and backpropagation, gradient descent, activation functions, Sequential API, Functional API, wide and deep models, Subclassing API for dynamic models, hyperparameter tuning, improved neural networks, vanishing/exploding gradient problem
- Non-saturating activation functions, optimization algorithms, gradient clipping, learning rate scheduling, regularization techniques, and transfer learning, End-to-end learning, recurrent neural networks (RNN), recurrent network architectures, long short-term memory (LSTM), gated recurrent unit (GRU), sequence models, convolutional neural networks (CNN), convolutional layers, powerful CNN models
- Object localization, object detection, segmentation, data augmentation, applications of CNN in transfer learning, and generative adversarial networks (GAN).
- Section 3. Signal Processing-Based Data Mining
An introduction to digital signal processing, analog-to-digital and digital-to-analog conversion (ADC & DAC), linear systems in signal processing, fundamentals of frequency domain analysis, power spectral density (PSD), and an introduction to filters.