Engineering Deep Learning
& Cybersecurity Systems
Curated by Ahmed BARGADY — a graduated Data Scientist, AI Engineer, and PhD researcher at UM6P. Combining first-principles mathematical derivations, production PyTorch/DGL code, and synthesis of research in APT detection via Graph Neural Networks.
AI & Deep Learning
Foundations to Frontiers
Cybersecurity & APT
Provenance Graph Analysis
Digital Twin & RAG
Agentic Systems & MCP
The Three-Step Mastery Framework
Every topic in this knowledge base is constructed with mathematical rigor and practical code.
Intuitive Conceptualization
High-level narrative explanations focused on the "Why" and real-world system intuition before diving into technical details.
Rigorous Mathematical Proofs
Step-by-step LaTeX formulations of objective loss functions, forward/backward equations, KKT conditions, and stochastic calculus.
Scratch Code Implementation
Vectorized, clean code written from scratch in Python, PyTorch, and DGL with line-by-line breakdown.
Featured APT Provenance Detection Papers
MAGIC (USENIX Security '24)
Masked Graph Autoencoder for Provenance APT Detection
ThreaTrace (IEEE TIFS '22)
Node-Level Provenance Graph Detection via GraphSAGE
UNICORN (NDSS '20)
Provenance Graph Embeddings via Graph Histograms & Hashing
Holmes (IEEE S&P '19)
Real-Time APT Detection via High-Level TTP Abstraction
Full Knowledge Matrix
Direct access to theoretical modules, datasets, and project architecture documents.