Robust Yet Efficient Conformal Prediction Sets

2024/07/12 by Soroush H. Zargarbashi, Zargarbashi, Soroush H., Mohammad Sadegh Akhondzadeh +3 · 6 citations
Computer Science · #Artificial Intelligence (cs.AI) #FOS: Computer and information sciences #Face and Expression Recognition #Machine Learning (cs.LG) #Neural Networks and Applications

paper · pdf · doi:10.48550/arxiv.2407.09165

openalex publication_date 2024/07/12 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Conformal prediction (CP) can convert any model's output into prediction sets guaranteed to include the true label with any user-specified probability. However, same as the model itself, CP is vulnerable to adversarial test examples (evasion) and perturbed calibration data (poisoning). We derive provably robust sets by bounding the worst-case change in conformity scores. Our tighter bounds lead to more efficient sets. We cover both continuous and discrete (sparse) data and our guarantees work both for evasion and poisoning attacks (on both features and labels).

Cited by

• Conformal Prediction in The Loop: A Feedback-Based Uncertainty Model for Trajectory Optimization
• Membership Inference Attacks with False Discovery Rate Control
• Enhancing Adversarial Robustness with Conformal Prediction: A Framework for Guaranteed Model Reliability
• One Sample is Enough to Make Conformal Prediction Robust
• Efficient Robust Conformal Prediction via Lipschitz-Bounded Networks
• Conformal Prediction with Corrupted Labels: Uncertain Imputation and Robust Re-weighting

Related

• Robust Conformal Prediction with a Single Binary Certificate
• Provably Reliable Conformal Prediction Sets in the Presence of Data Poisoning
• Provably Robust Conformal Prediction with Improved Efficiency
• Robust Conformal Prediction Using Privileged Information
• One Sample is Enough to Make Conformal Prediction Robust