paper-with-me

Papers

Test-time augmentation improves efficiency in conformal prediction

2025-05-28 · CVPR 2025 1 · Divya Shanmugam, Helen Lu, Swami Sankaranarayanan, John Guttag

A conformal classifier produces a set of predicted classes and provides a probabilistic guarantee that the set includes the true class. Unfortunately, it is often the case that conformal classifiers produce uninformatively large sets. In this work, we show that test-time augmentation (TTA)--a technique that introduces inductive biases during inference--reduces the size of the sets produced by conformal classifiers. Our approach is flexible, computationally efficient, and effective. It can be combined with any conformal score, requires no model retraining, and reduces prediction set sizes by 10%-14% on average. We conduct an evaluation of the approach spanning three datasets, three models, two established conformal scoring methods, different guarantee strengths, and several distribution shifts to show when and why test-time augmentation is a useful addition to the conformal pipeline.

📄 PDF Abstract BibTeX arXiv:2505.22764

Code (0)

등록된 구현이 없습니다.

Tasks

Conformal PredictionPrediction

Methods 이 논문이 사용한 방법론

SET Dynamic Sparse Training method where weight mask is updated randomly periodically

Similar Papers 제목 키워드 기반

Online Reasoning Calibration: Test-Time Training Enables Generalizable Conformal LLM Reasoning

2026-04-01 · Cai Zhou, Zekai Wang, Menghua Wu, Qianyu Julie Zhu 외 arxiv

While test-time scaling has enabled large language models to solve highly difficult tasks, state-of-the-art results come at exorbitant compute costs. These inefficiencies can be attributed to the miscalibration of post-t…

Annotation-Efficient Active Test-Time Adaptation with Conformal Prediction

2025-09-30 · Tingyu Shi, Fan Lyu, Shaoliang Peng arxiv

Active Test-Time Adaptation (ATTA) improves model robustness under domain shift by selectively querying human annotations at deployment, but existing methods use heuristic uncertainty measures and suffer from low data se…

Test-time Adaptation

Unified Conformalized Multiple Testing with Full Data Efficiency

2025-08-16 · Yuyang Huo, Xiaoyang Wu, Changliang Zou, Haojie Ren arxiv

Conformalized multiple testing offers a model-free way to control predictive uncertainty in decision-making. Existing methods typically use only part of the available data to build score functions tailored to specific se…

Conformal Prediction in Learning Under Privileged Information Paradigm with Applications in Drug Discovery

2018-03-29 · Niharika Gauraha, Lars Carlsson, Ola Spjuth

This paper explores conformal prediction in the learning under privileged information (LUPI) paradigm. We use the SVM+ realization of LUPI in an inductive conformal predictor, and apply it to the MNIST benchmark dataset …

Conformal PredictionDrug DiscoveryPredictionPrediction Intervals+1

Effect of Data Augmentation on Conformal Prediction for Diabetic Retinopathy

2025-08-19 · Rizwan Ahamed, Annahita Amireskandari, Joel Palko, Carol Laxson 외 arxiv

The clinical deployment of deep learning models for high-stakes tasks such as diabetic retinopathy (DR) grading requires demonstrable reliability. While models achieve high accuracy, their clinical utility is limited by …

Data Augmentation