Autodiff and Training Forensics

Repair a network with detached gradients, unstable initialization, and a misleading training curve. Validate gradients numerically and explain the repair.

Evidence and review

Deliver traces and a clean reproduction. Claims must trace to versioned code and artifacts. Screenshots, unexplained generated prose, and a single favorable run do not pass.

Oral defense

Explain the mechanism, identify the most dangerous assumption, interpret one failure from raw evidence, and name the condition that would reverse your decision.

Source backbone

Use Dive into Deep Learning, PyTorch documentation, primary papers, and Designing Data-Intensive Applications selectively.