Do not trust what you trust: Miscalibration in Semi-supervised Learning

• URL: http://arxiv.org/abs/2403.15567v1
• Date: Fri, 22 Mar 2024 18:43:46 GMT
• Title: Do not trust what you trust: Miscalibration in Semi-supervised Learning
• Authors: Shambhavi Mishra, Balamurali Murugesan, Ismail Ben Ayed, Marco Pedersoli, Jose Dolz,
• Abstract summary: State-of-the-art semi-supervised learning (SSL) approaches rely on highly confident predictions to serve as pseudo-labels that guide the training on unlabeled samples. We show that SSL methods based on pseudo-labels are significantly miscalibrated, and formally demonstrate the minimization of the min-entropy. We integrate a simple penalty term, which enforces the logit of the predictions on unlabeled samples to remain low, preventing the network predictions to become overconfident.
• Score: 21.20806568508201
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: State-of-the-art semi-supervised learning (SSL) approaches rely on highly confident predictions to serve as pseudo-labels that guide the training on unlabeled samples. An inherent drawback of this strategy stems from the quality of the uncertainty estimates, as pseudo-labels are filtered only based on their degree of uncertainty, regardless of the correctness of their predictions. Thus, assessing and enhancing the uncertainty of network predictions is of paramount importance in the pseudo-labeling process. In this work, we empirically demonstrate that SSL methods based on pseudo-labels are significantly miscalibrated, and formally demonstrate the minimization of the min-entropy, a lower bound of the Shannon entropy, as a potential cause for miscalibration. To alleviate this issue, we integrate a simple penalty term, which enforces the logit distances of the predictions on unlabeled samples to remain low, preventing the network predictions to become overconfident. Comprehensive experiments on a variety of SSL image classification benchmarks demonstrate that the proposed solution systematically improves the calibration performance of relevant SSL models, while also enhancing their discriminative power, being an appealing addition to tackle SSL tasks.

Related papers

• Towards the Mitigation of Confirmation Bias in Semi-supervised Learning: a Debiased Training Perspective [6.164100243945264]
  Semi-supervised learning (SSL) commonly exhibits confirmation bias, where models disproportionately favor certain classes. We introduce TaMatch, a unified framework for debiased training in SSL. We show that TaMatch significantly outperforms existing state-of-the-art methods across a range of challenging image classification tasks.
  arXiv  Detail & Related papers  (2024-09-26T21:50:30Z)
• Learning Label Refinement and Threshold Adjustment for Imbalanced Semi-Supervised Learning [6.904448748214652]
  Semi-supervised learning algorithms struggle to perform well when exposed to imbalanced training data. We introduce SEmi-supervised learning with pseudo-label optimization based on VALidation data (SEVAL) SEVAL adapts to specific tasks with improved pseudo-labels accuracy and ensures pseudo-labels correctness on a per-class basis.
  arXiv  Detail & Related papers  (2024-07-07T13:46:22Z)
• Generalized Semi-Supervised Learning via Self-Supervised Feature Adaptation [87.17768598044427]
  Traditional semi-supervised learning assumes that the feature distributions of labeled and unlabeled data are consistent. We propose Self-Supervised Feature Adaptation (SSFA), a generic framework for improving SSL performance when labeled and unlabeled data come from different distributions. Our proposed SSFA is applicable to various pseudo-label-based SSL learners and significantly improves performance in labeled, unlabeled, and even unseen distributions.
  arXiv  Detail & Related papers  (2024-05-31T03:13:45Z)
• A Channel-ensemble Approach: Unbiased and Low-variance Pseudo-labels is Critical for Semi-supervised Classification [61.473485511491795]
  Semi-supervised learning (SSL) is a practical challenge in computer vision. Pseudo-label (PL) methods, e.g., FixMatch and FreeMatch, obtain the State Of The Art (SOTA) performances in SSL. We propose a lightweight channel-based ensemble method to consolidate multiple inferior PLs into the theoretically guaranteed unbiased and low-variance one.
  arXiv  Detail & Related papers  (2024-03-27T09:49:37Z)
• InPL: Pseudo-labeling the Inliers First for Imbalanced Semi-supervised Learning [34.062061310242385]
  We present a new perspective of pseudo-labeling for imbalanced semi-supervised learning (SSL) We measure whether an unlabeled sample is likely to be in-distribution'' or out-of-distribution'' Experiments demonstrate that our energy-based pseudo-labeling method, textbfInPL, significantly outperforms confidence-based methods on imbalanced SSL benchmarks.
  arXiv  Detail & Related papers  (2023-03-13T16:45:41Z)
• Uncertainty Estimation by Fisher Information-based Evidential Deep Learning [61.94125052118442]
  Uncertainty estimation is a key factor that makes deep learning reliable in practical applications. We propose a novel method, Fisher Information-based Evidential Deep Learning ($mathcalI$-EDL) In particular, we introduce Fisher Information Matrix (FIM) to measure the informativeness of evidence carried by each sample, according to which we can dynamically reweight the objective loss terms to make the network more focused on the representation learning of uncertain classes.
  arXiv  Detail & Related papers  (2023-03-03T16:12:59Z)
• Complementing Semi-Supervised Learning with Uncertainty Quantification [6.612035830987296]
  We propose a novel unsupervised uncertainty-aware objective that relies on aleatoric and epistemic uncertainty quantification. Our results outperform the state-of-the-art results on complex datasets such as CIFAR-100 and Mini-ImageNet.
  arXiv  Detail & Related papers  (2022-07-22T00:15:02Z)
• Uncertainty-aware Pseudo-label Selection for Positive-Unlabeled Learning [10.014356492742074]
  We propose to tackle the issues of imbalanced datasets and model calibration in a positive-unlabeled learning setting. By boosting the signal from the minority class, pseudo-labeling expands the labeled dataset with new samples from the unlabeled set. Within a series of experiments, PUUPL yields substantial performance gains in highly imbalanced settings.
  arXiv  Detail & Related papers  (2022-01-31T12:55:47Z)
• Taming Overconfident Prediction on Unlabeled Data from Hindsight [50.9088560433925]
  Minimizing prediction uncertainty on unlabeled data is a key factor to achieve good performance in semi-supervised learning. This paper proposes a dual mechanism, named ADaptive Sharpening (ADS), which first applies a soft-threshold to adaptively mask out determinate and negligible predictions. ADS significantly improves the state-of-the-art SSL methods by making it a plug-in.
  arXiv  Detail & Related papers  (2021-12-15T15:17:02Z)
• In Defense of Pseudo-Labeling: An Uncertainty-Aware Pseudo-label Selection Framework for Semi-Supervised Learning [53.1047775185362]
  Pseudo-labeling (PL) is a general SSL approach that does not have this constraint but performs relatively poorly in its original formulation. We argue that PL underperforms due to the erroneous high confidence predictions from poorly calibrated models. We propose an uncertainty-aware pseudo-label selection (UPS) framework which improves pseudo labeling accuracy by drastically reducing the amount of noise encountered in the training process.
  arXiv  Detail & Related papers  (2021-01-15T23:29:57Z)
• Distribution Aligning Refinery of Pseudo-label for Imbalanced Semi-supervised Learning [126.31716228319902]
  We develop Distribution Aligning Refinery of Pseudo-label (DARP) algorithm. We show that DARP is provably and efficiently compatible with state-of-the-art SSL schemes.
  arXiv  Detail & Related papers  (2020-07-17T09:16:05Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.