Controlling the Probability of Unknown Unknowns in Pre-Deployment Testing of AI Systems
cs.AI, cs.CY, math.ST, stat.TH
Submitted: 2026-11-11
Updated: 2026-11-11
Code: https://github.com/rachelli2019/unseen-failures-code
Terminology
Sources
- Concrete Problems in AI Safety
- SWE-chat: Coding Agent Interactions From Real Users in the Wild
- Concentration inequalities in the infinite urn scheme for occupancy counts and the missing mass, with applications
- Optimal distribution-free sequential change detection with conformal martingales
- Confidence intervals for maximum unseen probabilities, with application to sequential sampling design
- Besting Good--Turing: Optimality of Non-Parametric Maximum Likelihood for Distribution Estimation
- Conformal changepoint localization
- Weighted-threshold Coupon Collection
- Blind-Spot Mass: A Good-Turing Framework for Quantifying Deployment Coverage Risk in Machine Learning Systems
- Red Teaming Language Models with Language Models
- Asymptotically optimal sequential change detection for bounded means
- E-valuator: Reliable Agent Verifiers with Sequential Hypothesis Testing
- Distribution-free changepoint localization after sequential change detection
- Model evaluation for extreme risks
- Rigorous Agent Evaluation: An Adversarial Approach to Uncover Catastrophic Failures
- Anytime Safe PAC Efficient Reasoning
Related papers
- Multi-Objective Structured Pruning of LLMs for Latency and Model Size Optimization
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- Embeddings for Preferences, Not Semantics
- Can We Trust Item Response Theory for AI Evaluation?
- Diagnosing Task Insensitivity in Language Agents