Papers/2608.13601
🧪 Test?View on arXiv

Hard Cases, Bad Labels: Testing Error Exposure and Error Location in Uncertainty Sampling Under Bounded Label Noise

Not provided in the abstract

active learninglabel noiseuncertainty samplingbinary classification
2608.13601
Builder Relevance
70%
1h ago

Abstract

This study investigates the performance of uncertainty sampling in active learning under various noise conditions and its impact on labeling efficiency.

Reality Card

Core Claim

Uncertainty sampling improved normalized balanced-accuracy area under the learning curve by 1.09 to 1.77 percentage points on all datasets under clean labels.

Method / Result

Uncertainty sampling showed a balanced accuracy improvement across datasets, with specific metrics indicating varying robustness depending on dataset and noise structure.

Limitations

The robustness of uncertainty sampling is dependent on dataset, budget, noise structure, and evaluation metric, which may limit generalizability.

Paper to code

Verified implementation resources so builders can test the paper’s claims instead of stopping at the abstract.

No verified implementation link has been attached yet. AIBuzzHub will keep this panel separate from unverified search results.
← Back to all papers