T-TF-028-007 AI Retraining Report
Table of contents
Purpose
Describe the purpose of this retraining report. Identify the predicate algorithm being retrained and the goals of the retraining effort.
Scope
Define the scope of the retraining, including which models are affected.
Related Documents
| Document ID | Title | Relationship |
|---|---|---|
| R-TF-028-001 | AI Description | |
| R-TF-028-002 | AI Development Plan | |
| R-TF-028-005 | AI Development Report | |
| R-TF-028-006 | AI Release Report (Previous) |
Predicate Algorithm Information
Describe the predicate algorithm being retrained.
| Item | Value |
|---|---|
| Algorithm Name | |
| Version | |
| Release Date | |
| Performance |
Modifications and Rationale
Description of Changes
Describe the specific changes made to the predicate algorithm.
Rationale
Explain the rationale for the retraining effort.
Change Classification
Classify the change as Minor or Major per GP-028 versioning scheme.
| Criterion | Assessment |
|---|---|
| Architecture Change | |
| Training Data Modification | |
| Hyperparameter Change | |
| Input/Output Change |
Change Classification: [Minor / Major]
Data Management
Dataset Modifications
Describe any modifications to the training dataset.
| Modification Type | Description | Rationale |
|---|---|---|
| Images Added | ||
| Images Removed | ||
| Labels Fixed |
Dataset Statistics
Before Retraining:
| Item | Value |
|---|---|
| Total Images | |
| Categories |
After Retraining:
| Item | Value |
|---|---|
| Total Images | |
| Categories |
Data Quality Verification
Describe the data quality verification performed.
Training Process
Training Configuration
Describe any changes to the training configuration.
| Parameter | Predicate | Retrained |
|---|---|---|
| Architecture | ||
| Learning Rate | ||
| Epochs | ||
| Data Augmentation |
Training Summary
Summarize the training process.
Performance Testing
Test Methodology
Describe the test methodology. Confirm use of the same test data and metrics as the predicate algorithm.
Performance Comparison
Compare performance between predicate and retrained algorithms.
| Metric | Predicate | Retrained | Success Criterion | Outcome |
|---|---|---|---|---|
Non-Regression Analysis
Demonstrate that the retrained algorithm shows non-regression in performance.
Subgroup Analysis
Provide performance comparison across relevant subgroups.
| Subgroup | Predicate | Retrained |
|---|---|---|
Risk Assessment
Risk Identification
Identify any new risks introduced by the retraining.
Risk Evaluation
Evaluate the impact of the retraining on existing risks.
Benefit-Risk Assessment
Provide the benefit-risk assessment for this retraining.
Conclusions
State the conclusions from the retraining effort.
Release Information
Reference the new AI Release Report (R-TF-028-006) if applicable.
| Item | Value |
|---|---|
| New Version | |
| Release Date |
Related Documents
| Document ID | Title |
|---|---|
| R-TF-028-006 | |
| R-TF-028-010 | |
| R-TF-028-011 |
Signature meaning
The signatures for the approval process of this document can be found in the verified commits at the repository for the QMS. As a reference, the team members who are expected to participate in this document and their roles in the approval process, as defined in Annex I Responsibility Matrix of the GP-001, are:
- Author: JD-009
- Reviewer: JD-009
- Approver: JD-005