Single Trajectory Conformal Prediction
Not on the map. Why?
Abstract (arXiv)
We study the performance of risk-controlling prediction sets (RCPS), an empirical risk minimization-based formulation of conformal prediction, with a single trajectory of temporally correlated data from an unknown stochastic dynamical system. First, we use the blocking technique to show that RCPS attains performance guarantees similar to those enjoyed in the iid setting whenever data is generated by asymptotically stationary and contractive dynamics. Next, we use the decoupling technique to characterize the graceful degradation in RCPS guarantees when the data generating process deviates from stationarity and contractivity. We conclude by discussing how these tools could be used toward a unified analysis of online and offline conformal prediction algorithms, which are currently treated with very different tools.
Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 32 of its 48 references that are in the map. First posted on arXiv 3 Jun 2024.
Builds on
32 of 48 references.
References · 32
- 2016Distribution-Free Predictive Inference For Regressioncited by 657
- 2021Adaptive Conformal Inference Under Distribution Shiftcited by 360
- 2022Conformal prediction beyond exchangeabilitycited by 348
- 2022Conformal Risk Controlcited by 290
- 2021Distribution-Free, Risk-Controlling Prediction Setscited by 278
- 2018Learning Without Mixing: Towards A Sharp Analysis of Linear System Identificationcited by 158
- 2023Conformal Prediction With Conditional Guaranteescited by 146
- 2022Conformal Inference for Online Prediction with Arbitrary Distribution Shiftscited by 123
- 2022Safe Planning in Dynamic Environments using Conformal Predictioncited by 112
- 2019Nested conformal prediction and quantile out-of-bag ensemble methodscited by 88
- 2023Improved Online Conformal Prediction via Strongly Adaptive Online Learningcited by 76
- 2018Exact and Robust Conformal Inference Methods for Predictive Machine Learning With Dependent Datacited by 64
20 more
- 2020Conformal prediction for time seriescited by 53
- 2024Online conformal prediction with decaying step sizescited by 53
- 2022Statistical Learning Theory for Control: A Finite Sample Perspectivecited by 52
- 2023How to Trust Your Diffusion Model: A Convex Optimization Approach to Conformal Risk Controlcited by 50
- 2024Conformal prediction for multi-dimensional time series by ellipsoidal setscited by 37
- 2022Learning with little mixingcited by 33
- 2023Conformal Prediction Regions for Time Series using Linear Complementarity Programmingcited by 33
- 2018On the optimality of the Hedge algorithm in the stochastic regimecited by 32
- 2023Non-Exchangeable Conformal Risk Controlcited by 29
- 2017Least-Squares Temporal Difference Learning for the Linear Quadratic Regulatorcited by 27
- 2022Single Trajectory Nonparametric Learning of Nonlinear Dynamicscited by 26
- 2024Conformal Validity Guarantees Exist for Any Data Distribution (and How to Find Them)cited by 26
- 2024Localized Adaptive Risk Controlcited by 16
- 2023Physics Constrained Motion Prediction with Uncertainty Quantificationcited by 11
- 2023Data-Driven Reachability Analysis of Stochastic Dynamical Systems with Conformal Inferencecited by 9
- 2020Relaxing the I.I.D. Assumption: Adaptively Minimax Optimal Regret via Root-Entropic Regularizationcited by 8
- 2022Extending Conformal Prediction to Hidden Markov Models with Exact Validity via de Finetti's Theorem for Markov Chainscited by 8
- 2023Randomization Tests for Adaptively Collected Datacited by 7
- 2023The noise level in linear regression with dependent datacited by 6
- 2023PAC-Bayes Generalization Certificates for Learned Inductive Conformal Predictioncited by 6
Built on by
Citing papers per year: 2024: 1.
1 paper.