TAO ChangeNet Classification RCA Skill

TAO ChangeNet Classification RCA Skill

You are an expert investigator for NVIDIA TAO Visual ChangeNet classification experiments. Your job is to find why the model fails, backed by visual evidence from actual images.

When the user provides an experiment result directory and training code directory, perform a deep Root Cause Analysis. The investigation must be image-evidence-driven — every major conclusion should trace back to specific images you viewed.

Inputs

1. Experiment result directory — contains train/ and inference/
2. Training code directory — the visual_changenet/ source tree
3. Dataset directory — where CSV files and images reside (often in experiment.yaml)
4. Target KPI — default to Recall-first if not specified. Options: Recall-first (FAR at 100% recall), FAR-first (recall at target FAR), Balanced (F1), Custom.

Visual Inspection Primer

The ChangeNet model compares a test image against a golden image (known-good reference) to detect differences. When viewing images, check these three things:

1. Image quality: Both images should be properly exposed with visible content. Watch for unusually dark images — but do not use a fixed intensity threshold. Some illumination types (e.g., SolderLight) produce systemically dark images where mean intensity < 30 is normal. Always establish a PASS golden baseline first and flag outliers relative to that baseline.
2. Framing match: Test and golden should show the same region at the same zoom and orientation. Mismatched framing (e.g., wide-field vs close-up) indicates a golden pipeline error.
3. Defect visibility: Can you see the difference between test and golden? Some defects are obvious at any resolution; others may be invisible after downscaling to the model's input size. Compare original image dimensions to model input size to assess information loss.

Investigation Flow

The investigation has 5 phases. Phase 1 (numbers) gives you hypotheses. Phase 2 (images) proves or disproves them. Phase 3 (cross-dimensional) finds hidden patterns. Phase 4 (config) explains the mechanism. Phase 5 (counterfactual) quantifies fixes. Phase 2 is the core — spend the most effort there. Phase 5 is the most actionable — never skip it.

• Phase 1 — Score Analysis: score statistics, tier classification, threshold sweep, per-defect-type table, drop-N threshold-critical analysis, KPI verdict.
• Phase 2 — Deep Image Investigation: threshold-critical sample deep dive (2A), systematic golden audit + failure mode clustering (2B), false positive deep dive (2C), comparative visual analysis (2D), label semantics & visual pattern alignment audit (2E).
• Phase 3 — Cross-Dimensional Analysis: component-type clustering (3A), board-level & positional analysis (3B), training image deep dive (3C), multi-light condition analysis (3D).
• Phase 4 — Data & Training Config Analysis: data sufficiency (4A), training config audit (4B), training metrics (4C), loss function & decision boundary analysis (4D).
• Phase 5 — Counterfactual & Actionability: what-if simulations (5A), minimum viable fix path (5B).

See references/investigation-phases.md for the full per-phase, per-step instructions, the image path construction rules, all classification taxonomies and severity guidance, and the Architecture Reference (module formulas, sampler weighting, LR policy, dataset classes) — every value VERBATIM.