Atlas · GenAI 2026

Deep Learning & Foundation Model Architectures

33 skills · ontology graph below shows relations within this section.

What this domain covers

This edition groups 33 capabilities in Deep Learning & Foundation Model Architectures across 12 named categories. The inventory contains 23 concepts and 10 tools; 19 skills appeared in at least two of the three original research runs. The remaining entries stay visible with their lower coverage so a reader can distinguish taxonomy scope from research-system agreement.

Current category labels: DL Frameworks · Deep Learning Fundamentals · Efficient & Small Models · Foundation Model Ecosystem · Generative Architectures · Graph Neural Networks · Multimodal Architectures · Neural Architectures · and 4 more

Frequent learning foundations

  1. Deep Learning supports 7 mapped skills
  2. Transformer Architecture supports 5 mapped skills
  3. Linear Algebra supports 4 mapped skills
  4. Diffusion Models supports 2 mapped skills
  5. Convolutional Neural Networks supports 1 mapped skill