-TensorFlow Basics
-MLP, Learning rate, Overfitting, and Hyper-parameters(In this task you will implement an MLP model for virtual sensing using the flood dataset. The objectives are: Implementing an MLP model via TensorFlow Functional API. Getting more familiar with model fitting and overfitting. Implementing early stopping. Exploring hyperparameters and their influence. Selecting model architecture.)
-CNNs, ResNets, Transfer Learning and Grad-CAM( - Work with a dataset consisting of images of electronic waste.
- Implement a CNN for classification
- Compare MLP and CNN
- Vary architecture to improve model performance
- Implement a ResNet
- Apply transfer learning using ResNet50
- Apply Grad-CAM class activation visualization)
-Anomaly Detection using Autoencoders (- Work with a dataset for dynamic behavior analysis of automotive software systems.
- Implement autoencoders (AEs) for unsupervised anomaly detection.
- Compare MLP-based and CNN-based AEs.
- Vary architecture to improve model performance.
- Apply scheduled learning rate.
- Evaluate the performance of the model.)