Pith. sign in

REVIEW 6 cited by

Leave-One-Out Stable Conformal Prediction

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2504.12189 v1 pith:LOQZILJZ submitted 2025-04-16 stat.ML cs.LG

Leave-One-Out Stable Conformal Prediction

classification stat.ML cs.LG
keywords methodconformalpredictionstabilityleave-one-outcompareddatastable
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Conformal prediction (CP) is an important tool for distribution-free predictive uncertainty quantification. Yet, a major challenge is to balance computational efficiency and prediction accuracy, particularly for multiple predictions. We propose Leave-One-Out Stable Conformal Prediction (LOO-StabCP), a novel method to speed up full conformal using algorithmic stability without sample splitting. By leveraging leave-one-out stability, our method is much faster in handling a large number of prediction requests compared to existing method RO-StabCP based on replace-one stability. We derived stability bounds for several popular machine learning tools: regularized loss minimization (RLM) and stochastic gradient descent (SGD), as well as kernel method, neural networks and bagging. Our method is theoretically justified and demonstrates superior numerical performance on synthetic and real-world data. We applied our method to a screening problem, where its effective exploitation of training data led to improved test power compared to state-of-the-art method based on split conformal.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 6 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. How Many Iterations to Jailbreak? Dynamic Budget Allocation for Multi-Turn LLM Evaluation

    cs.LG 2026-05 unverdicted novelty 7.0

    DAPRO provides the first dynamic, theoretically guaranteed way to allocate interaction budgets across test cases for bounding time-to-event in multi-turn LLM evaluations, achieving tighter coverage than static conform...

  2. How Many Iterations to Jailbreak? Dynamic Budget Allocation for Multi-Turn LLM Evaluation

    cs.LG 2026-05 conditional novelty 7.0

    DAPRO adaptively spends a fixed interaction budget across multi-turn LLM conversations and builds valid lower bounds on time-to-jailbreak, with coverage error scaling as the square root of the mean censoring weight.

  3. Classification-Powered Conformal Inference for Zero-inflated Outcomes

    stat.ME 2026-05 unverdicted novelty 7.0

    A classification-integrated conformal framework for zero-inflated outcomes that guarantees marginal coverage and asymptotic minimal length under exchangeability, independent of the underlying models.

  4. How Many Iterations to Jailbreak? Dynamic Budget Allocation for Multi-Turn LLM Evaluation

    cs.LG 2026-05 unverdicted novelty 6.5

    DAPRO dynamically allocates multi-turn LLM evaluation budget to produce valid, tighter lower bounds on iterations-to-event without the conditional-independence assumption of prior conformal survival methods.

  5. Approximating full conformal prediction: distribution free guarantees via the tournament correction

    stat.ME 2026-05 unverdicted novelty 6.0

    Tournament corrections approximate full conformal prediction while providing rigorous marginal coverage of 1-2α (tightening to ~1-α under stability).

  6. Approximating full conformal prediction: distribution free guarantees via the tournament correction

    stat.ME 2026-05 unverdicted novelty 6.0

    Introduces a tournament correction framework for approximating full conformal prediction with rigorous 1-2α marginal coverage guarantees that generalize leave-one-out cross-conformal prediction.