research area
Responsible AI
Interpretability is only useful if it changes how systems are built. We work on robustness under adversarial and natural shift, fairness auditing, data quality, and reporting standards that make a model’s limitations legible before deployment, not after.
Projects
responsible-ai
responsible-ai
FairSight
A dataset auditing tool that surfaces demographic imbalance and labeling inconsistency before training begins.
Publications
Auditing Large-Scale Vision Datasets for Demographic Fairness
Priya Natarajan, Dev Anand · FAccT · 2025
A Benchmark for Failure-Mode Detection Under Distribution Shift
Dev Anand, Mira Solheim · ICCV · 2025
Adversarial Robustness in Multi-Object Detection
Priya Natarajan, Mira Solheim · CVPR · 2024
Concept Probes as Transparent Reporting for Deployed Models
Mira Solheim, Sana Idris · NeurIPS · 2024