AI-Projects-CV-NLP

These are the assignments of my Advanced Machine Learning course at AUB. Here is a small description for each one.

1. Simple Linear and Logistic Regression with L2 regularization.
2. Feed Forward Network for image classification.
3. CNN implementation.
4. Computer Vision:

• Object Detection: Fine-tuned Deeplabv3_RESNet model and performed Image Segmentation on a flower dataset based on IoU loss.
• Face Recognition: Created triplets of images to train Siamese networks, used Xception pre-trained model and applied transfer learning. Performed Face verification and recognition for anchor images.
• Style Transfer: Used VGGNet pre-trained model. Computed content loss at one convolution layer and style loss using gram matrices at all convolution layers.

5. Image Captioning: Input is a feature-map of an image reduced using PCA (encoder), and the output is a caption describing the image. Implemented forward and backward propagation of RNN, LSTM and word embedding.
6. LLM completion: Completed the architecture of BERT... and applied it for NLP tasks (named entity recognition, natural language inference and sentiment analysis).

Final Project: Deep Learning Pipeline for Lung Cancer Diagnosis (not shared due to confidentiality).

• Implemented computer vision models for lung cancer classification and segmentation such as ViT, YOLO, UNet, ResNet, MobileNet ...