IBR Identification and Learned-Model Validation
Configuration Uncertainty and Fair Baselines
T1-03 · Module outline
Evaluate models without giving one method hidden information advantages.
Learning objectives
- Separate configuration uncertainty from parameter uncertainty.
- Match candidate knowledge, initial conditions, and budgets.
Concept outline
- GFL, GFM, and parallel configurations
- Oracle diagnostics and validation-based selection
- Noise, seeds, failures, and paired comparisons
Experiment direction · Planned
Simulation and code experiment
Write a protocol for the fixed low-frequency configuration models.
Model explanations, parameter settings, runnable code, and result interpretation will be developed here.
Practice and discussion
Identify one unfair baseline and repair its information budget.
Worked solutions and feedback will accompany the full lesson.
Material preparation
Related research cases or review records are available and need adaptation into a lesson with a declared scope.
Develop the planned configuration-learning experiments before presenting new results.