WG2: Transparent vs Black Box Models

Working Group 2: Transparent versus Black Box Decision-Support Models#

Leader: Prof. Petre Lameski (North Macedonia) Co-Leader: Dr. Kristina Sutiene

Mission#

WG2 investigates the trade-offs between model performance and interpretability, developing methods to make complex financial models more transparent without sacrificing accuracy.

Focus Areas#

Explainable AI (XAI) in Finance#

SHAP and LIME implementations
Model-agnostic explanations
Local and global interpretability

Credit Scoring Models#

Fair lending analysis
Bias detection and mitigation
Regulatory compliance

Risk Assessment#

Stress testing methodologies
Scenario analysis
Model validation

Algorithmic Trading#

Strategy transparency
Performance attribution
Regulatory reporting

Model Interpretability#

Feature engineering
Model simplification
Hybrid approaches

Key Deliverables#

XAI implementation guidelines
Credit scoring transparency toolkit
Stress test design frameworks
Best practice documents

Members#

WG2 includes approximately 130 researchers from across the network.