AINS graduate course

AINS6003 · Deep Learning & Neural Networks

Core sequence · catalog 2026–27

Description

Neural architectures, training at scale, and practical debugging for vision, sequence, and representation learning tasks.

Course shells on the Castalia LMS are provisioned per license; this link opens the LMS to explore the guest demo or landing experience.

Open Castalia LMS Back to catalog

Buy license Continue on the purchase hub to request a license or institutional quote.

Syllabus outline

  1. Modules 1–2 · Networks & training

    • Backpropagation and autodiff mental models
    • Optimization and learning-rate strategies
    • Initialization and normalization

  2. Modules 3–4 · Architectures

    • CNNs and spatial inductive bias
    • Sequence models and attention (intro)
    • Transfer learning workflows

  3. Modules 5–6 · Engineering

    • Compute budgets and mixed precision
    • Experiment tracking and reproducibility
    • Failure modes and robustness checks