Pinned Loading

1. Video_summarizer Video_summarizer Public

   Internship project at Proxym IT – AI-powered video summarizer that goes beyond transcripts (Detects and extracts code directly from frames using CLIP + OCR and frames histogram analysis)

   Jupyter Notebook

2. Fire-Detection-and-Localization-using-CLIP-DINOv2 Fire-Detection-and-Localization-using-CLIP-DINOv2 Public

   A two-stage pipeline for detecting and localizing forest fires using CLIP for zero-shot classification(classes :fire /no fire) and DINOv2 for patch-level anomaly localization(where is the fire).

   Jupyter Notebook

3. EEG-data-augmentation EEG-data-augmentation Public

   GAN /VAE /DAE potential for data augmentation for EEG data / Arabs_in_neuroscience community

   Jupyter Notebook 1

4. Self-Supervised-Learning-fo-Few-Shot-Botnet-Detection Self-Supervised-Learning-fo-Few-Shot-Botnet-Detection Public

   Self-supervised learning pipeline combining Denoising Autoencoders and BYOL for robust botnet detection in noisy network traffic data

   Jupyter Notebook 1

5. BYOL-SimCLR-for-Flow-Embeddings-for-Botnet-Detection BYOL-SimCLR-for-Flow-Embeddings-for-Botnet-Detection Public

   Self-supervised learning on network flow data using SimCLR ( BYOL to add later too) to learn embeddings that can be used for few-shot botnet classification with CTU13 dataset

   Jupyter Notebook 2

6. Challenge Challenge Public

   Liar Nodes is an adversarial node classification challenge on single-cell cancer graphs. Node features (gene expression embeddings) and graph connectivity are partially corrupted, creating structur…

   Jupyter Notebook 17