SEED-SET: Scalable Evolving Experimental Design for System-level Ethical Testing

TL;DR

SEED-SET is a Bayesian framework for ethically benchmarking autonomous systems, combining objective metrics and stakeholder preferences to efficiently identify high-quality test cases in complex environments.

Contribution

It introduces a hierarchical Gaussian Process model and a novel acquisition strategy for ethical testing, addressing the lack of evaluation metrics and stakeholder subjectivity.

Findings

Outperforms baseline methods in test candidate quality.

Generates up to 2x more optimal test cases.

Improves coverage of high-dimensional search spaces by 1.25x.

Abstract

As autonomous systems such as drones, become increasingly deployed in high-stakes, human-centric domains, it is critical to evaluate the ethical alignment since failure to do so imposes imminent danger to human lives, and long term bias in decision-making. Automated ethical benchmarking of these systems is understudied due to the lack of ubiquitous, well-defined metrics for evaluation, and stakeholder-specific subjectivity, which cannot be modeled analytically. To address these challenges, we propose SEED-SET, a Bayesian experimental design framework that incorporates domain-specific objective evaluations, and subjective value judgments from stakeholders. SEED-SET models both evaluation types separately with hierarchical Gaussian Processes, and uses a novel acquisition strategy to propose interesting test candidates based on learnt qualitative preferences and objectives that align with the stakeholder preferences. We validate our approach for ethical benchmarking of autonomous agents on two applications and find our method to perform the best. Our method provides an interpretable and efficient trade-off between exploration and exploitation, by generating up to $2\times$ optimal test candidates compared to baselines, with $1.25\times$ improvement in coverage of high dimensional search spaces.