Contrastive Explanation Method

Description

The Contrastive Explanation Method (CEM) explains model decisions by generating contrastive examples that reveal what makes a prediction distinctive. It identifies 'pertinent negatives' (minimal features that could be removed to change the prediction) and 'pertinent positives' (minimal features that must be present to maintain the prediction). This approach helps users understand not just what led to a decision, but what would need to change to achieve a different outcome, providing actionable insights for decision-making.

Example Use Cases

Explainability

Explaining loan application rejections by showing that removing recent late payments (pertinent negative) or adding £5,000 more annual income (pertinent positive) would change the decision to approval, giving applicants clear actionable guidance.

Analysing medical diagnosis models by identifying that removing a specific symptom combination would change a high-risk classification to low-risk, helping clinicians understand the critical diagnostic factors.

Transparency

Providing transparent hiring decisions by showing job candidates exactly which qualifications (pertinent positives) they need to acquire or which application elements (pertinent negatives) might be hindering their success.

Limitations

Computationally expensive as it requires solving an optimisation problem for each individual instance to find minimal perturbations.
Results can be highly sensitive to hyperparameter settings, requiring careful tuning to produce meaningful explanations.
May generate unrealistic or impossible contrastive examples if constraints are not properly specified, leading to impractical recommendations.
Limited to scenarios where feature perturbations are meaningful and actionable, making it less suitable for immutable characteristics or highly constrained domains.

Resources

Explanations based on the Missing: Towards Contrastive Explanations with Pertinent Negatives

Research Paper•Amit Dhurandhar et al.•Feb 21, 2018

Interpretable Machine Learning

Documentation

Benchmarking and survey of explanation methods for black box models

Documentation•Bodria Francesco et al.•Jan 1, 2023

Related Techniques

Name	Description	Assurance Goals
Exponentiated Gradient Reduction	An in-processing fairness technique based on Agarwal et al.'s reductions approach that transforms fair classification into a sequence of cost-sensitive classification problems. The method uses an exponentiated gradient algorithm to iteratively reweight training data, returning a randomised classifier that achieves the lowest empirical error whilst satisfying fairness constraints. This reduction-based framework provides theoretical guarantees about both accuracy and constraint violation, making it suitable for various fairness criteria including demographic parity and equalised odds.	Fairness Transparency Reliability
Prejudice Remover Regulariser	An in-processing fairness technique that adds a fairness penalty to machine learning models to reduce bias against protected groups. The method works by minimising 'mutual information' - essentially reducing how much the model's predictions reveal about sensitive attributes like race or gender. By adding this penalty term to the learning objective (typically in logistic regression), the technique ensures predictions become less dependent on protected features. This addresses not only direct discrimination but also indirect bias through correlated features. Practitioners can adjust a tuning parameter to balance between maintaining accuracy and removing prejudice from the model.	Fairness Transparency Reliability
Data Version Control	Data Version Control (DVC) is a Git-like version control system specifically designed for machine learning data, models, and experiments. It tracks changes to large data files, maintains reproducible ML pipelines, and creates a complete audit trail of data transformations, model training, and evaluation processes. DVC works alongside Git to provide end-to-end lineage tracking from raw data through preprocessing, training, and deployment, enabling teams to reproduce any model version and understand exactly how datasets evolved throughout the ML lifecycle.	Transparency Reliability
Runtime Monitoring and Circuit Breakers	Runtime monitoring and circuit breakers establish continuous surveillance of AI/ML systems in production, tracking critical metrics such as prediction accuracy, response times, input characteristics, output distributions, and system resource usage. When monitored parameters exceed predefined safety thresholds or exhibit anomalous patterns, automated circuit breakers immediately trigger protective actions including request throttling, service degradation, system shutdown, or failover to backup mechanisms. This approach provides real-time defensive capabilities that prevent cascading failures, ensure consistent service reliability, and maintain transparent operation status for stakeholders monitoring system health.	Safety Reliability Transparency
Path-Specific Counterfactual Fairness Assessment	A causal fairness evaluation technique that assesses algorithmic discrimination by examining specific causal pathways in a model's decision-making process. Unlike general counterfactual fairness, this approach enables practitioners to identify and intervene on particular causal paths that may introduce bias whilst preserving other legitimate pathways. The method uses causal graphs to distinguish between direct discrimination (through protected attributes) and indirect discrimination (through seemingly neutral factors that correlate with protected attributes), allowing for more nuanced fairness assessments in complex causal settings.	Fairness Transparency Reliability
Layer-wise Relevance Propagation	Layer-wise Relevance Propagation (LRP) explains neural network predictions by working backwards through the network to show how much each input feature contributed to the final decision. It follows a simple conservation rule: the total contribution scores always add up to the original prediction. Starting from the output, LRP distributes 'relevance' backwards through each layer using different rules depending on the layer type. This creates a detailed breakdown showing which input features helped or hindered the prediction, making it easier to understand why the network made a particular decision.	Explainability Transparency