| name | KSB-D08-K0049 |
| description | Model Interpretability Requirements: Understanding of interpretability requirements for AI models in safety-critical pharmacovigilance ap... |
| version | 1.0.0 |
| domain | D08 |
| domain_name | Signal Detection & Management |
| type | Knowledge |
| proficiency_level | L3 |
| bloom_level | apply |
| triggers | ["explain model interpretability requirements","what is model interpretability requirements","apply model interpretability requirements"] |
| epa_mapping | EPA-04, EPA-05 |
| cpa_mapping | CPA-02 |
| regulatory_refs | |
KSB-D08-K0049: Model Interpretability Requirements
Overview
Domain: D08 - Signal Detection & Management
Type: Knowledge
Proficiency Level: L3 (Competent - Minimal supervision)
Bloom Level: Apply
Description
Understanding of interpretability requirements for AI models in safety-critical pharmacovigilance applications including distinction between inherently interpretable models (decision trees, linear models) and post-hoc interpretability methods, trade-offs between model complexity and interpretability, and regulatory expectations for model transparency. Knowledge of interpretability standards across FDA, EMA, and CIOMS frameworks.
Context
- Major Section: AI Integration and Advanced Analytics
- Section: AI Transparency and Explainability
EPA Mapping
- EPA-04:3009-3010
- EPA-05:3011-3012
CPA Pathway
Regulatory References
Instructions
When this skill is activated, Claude should:
- Demonstrate L3 proficiency in model interpretability requirements
- Apply apply level cognitive skills to apply the topic
- Reference relevant regulatory guidance (general PV standards)
- Connect to related EPAs: EPA-04, EPA-05
Key Competencies
- Understanding of interpretability requirements for AI models in safety-critical pharmacovigilance applications including distinction between inherently interpretable models (decision trees, linear models) and post-hoc interpretability methods, trade-offs between model complexity and interpretability, and regulatory expectations for model transparency. Knowledge of interpretability standards across FDA, EMA, and CIOMS frameworks.
Assessment Criteria
- Can apply core concepts independently
- Demonstrates understanding of regulatory context
- Applies knowledge appropriately to PV scenarios
Related Skills
- Other D08 skills in AI Transparency and Explainability
- Cross-domain integrations per DAG architecture
Generated from PV KSB Framework v1.0 | 2025-12-31
