Pith. sign in

REVIEW 7 cited by

SELF: Learning to Filter Noisy Labels with Self-Ensembling

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 1910.01842 v1 pith:FJYWP7ZS submitted 2019-10-04 cs.CV cs.LGstat.ML

classification cs.CVcs.LGstat.ML
keywords traininglabelslearningdifferentnetworknoisydatasetestimates
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Deep neural networks (DNNs) have been shown to over-fit a dataset when being trained with noisy labels for a long enough time. To overcome this problem, we present a simple and effective method self-ensemble label filtering (SELF) to progressively filter out the wrong labels during training. Our method improves the task performance by gradually allowing supervision only from the potentially non-noisy (clean) labels and stops learning on the filtered noisy labels. For the filtering, we form running averages of predictions over the entire training dataset using the network output at different training epochs. We show that these ensemble estimates yield more accurate identification of inconsistent predictions throughout training than the single estimates of the network at the most recent training epoch. While filtered samples are removed entirely from the supervised training loss, we dynamically leverage them via semi-supervised learning in the unsupervised loss. We demonstrate the positive effect of such an approach on various image classification tasks under both symmetric and asymmetric label noise and at different noise ratios. It substantially outperforms all previous works on noise-aware learning across different datasets and can be applied to a broad set of network architectures.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 7 Pith papers

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

  1. Tackling the Noisy Elephant in the Room: Label Noise-robust Out-of-Distribution Detection via Loss Correction and Low-rank Decomposition

    cs.LG 2025-09 reject novelty 6.0 of 10

    NOODLE corrects noisy labels with a transition matrix and cleans features via low-rank sparse decomposition, improving OOD detection under label noise.

  2. Commuting Distance Regularization for Timescale-Dependent Label Inconsistency in EEG Emotion Recognition

    cs.CV 2025-07 conditional novelty 6.0 of 10

    Graph-commute-distance regularizers LVL and LGCL improve aggregate rank of EEG emotion recognition under label inconsistency across three backbones on DREAMER and DEAP.

  3. Calibrating Pre-trained Language Classifiers on LLM-generated Noisy Labels via Iterative Refinement

    cs.CL 2025-05 conditional novelty 6.0 of 10

    SiDyP improves classifiers trained on LLM-generated noisy labels by retrieving likely true labels from embedding-space neighbors and iteratively refining them with a simplex diffusion model, reporting average gains of...

  4. MoMBS: Mixed-order minibatch sampling enhances model training from diverse-quality images

    cs.CV 2025-05 conditional novelty 6.0 of 10

    Pairing high-difficulty with low-difficulty images, ranked by loss and uncertainty, makes minibatch training more effective and improves accuracy on four computer vision tasks.

  5. Self-Boost via Optimal Retraining: An Analysis via Approximate Message Passing

    cs.LG 2025-05 conditional novelty 6.0 of 10

    For binary classification with noisy labels, the paper derives the Bayes-optimal function for combining a model's current predictions with the given labels during retraining, and shows a fitted version improves linear...

  6. Unleashing the Power of Large Language Model for Denoising Recommendation

    cs.IR 2025-02 conditional novelty 6.0 of 10

    LLaRD uses LLM-generated preference and relation knowledge plus an information-bottleneck objective to denoise implicit feedback and improve recommendation accuracy on Steam, Yelp, and Amazon-Book.

  7. Why Can Accurate Models Be Learned from Inaccurate Annotations?

    cs.LG 2025-05 conditional novelty 5.0 of 10

    Principal subspaces of classifier weights are largely preserved under moderate label inaccuracy, which explains why models still learn from noisy labels.

Pith tools