Proposes a new framework for learning from labeled and unlabeled data.
problem Learning from unlabeled and multi-label samples with arbitrary loss functions.
method Multi-complementary and unlabeled learning framework.
result Effective estimation of classification risk with optimal convergence rate.
Boosting with unlabeled data achieves optimal sample complexity in agnostic settings.
problem Boosting's sample inefficiency in agnostic learning.
method Designing an agnostic boosting algorithm with unlabeled data to match ERM's sample complexity.
result The total sample complexity is optimal, with a vanishing fraction needing to be labeled.
Generative model creates synthetic unlabeled data for SSL.
problem Training SSL models without real unlabeled datasets.
method Meta-optimized synthetic samples generated from generative models.
result Synthetic samples improve SSL performance more efficiently than real unlabeled data.
Improves PU learning for imbalanced data with practical AUL estimation and new training method.
problem Training binary classifiers on datasets with only positive and unlabeled samples.
method Asymptotic unbiased AUL estimation and ProbTagging for imbalanced data.
result ProbTagging increases AUC by up to 10% on industrial and artificial data sets.
DEDPUL improves PU learning by estimating proportions and classifying unlabeled data.
problem Analog to supervised binary classification with only positive samples clean and unlabeled mixtures of positive and negative.
method Applies a post-processing procedure to any classifier trained to distinguish positive and unlabeled data, estimating proportions alongside classification.
result Outperforms state-of-the-art in both proportion estimation and PU classification.
Enhances adversarial robustness with unlabeled out-of-domain data.
problem Improving robustness of models against adversarial attacks.
method Leveraging unlabeled data from multiple domains to bridge the sample complexity gap in adversarial robustness.
result Better adversarial robustness achieved when unlabeled data comes from a shifted domain.
This paper proposes a new method to use unlabeled data as positive data, improving PU learning.
problem Improving positive-unlabeled learning methods in deep learning.
method Labeling large-loss unlabeled data as positive data and developing a new learning objective.
result The proposed method outperforms the latest importance reweighting method in experiments.
Paper proposes a new softmax loss for better performance in Positive and Unlabeled data tasks.
problem Current softmax losses and sampling schemes have drawbacks in Positive and Unlabeled learning.
method Proposes Relaxed Softmax (RS) loss and a new negative sampling scheme.
result New training objective drives uplifts in performance on textual and recommendation datasets.
LoD improves model safety by integrating unlabeled wild data, reducing OOD misclassification.
problem Improving model safety and reliability using unlabeled wild data containing both in-distribution and out-of-distribution samples.
method Intentionally label-noisifying unlabeled wild data to enable joint learning of labeled ID and OOD data, distinguishing losses between ID and OOD samples.
result LoD framework achieves superior OOD detection without requiring thresholds, improving model safety.
Paper proposes a method to estimate class prior from positive and unlabeled data.
problem Estimating class prior in unlabeled datasets when labeled data is not available.
method Use penalized divergences to fit a mixture of class-wise distributions to the unlabeled data distribution.
result Correct estimation of class prior using only positive samples and penalized L1-distance. A semi-supervised learning method using predefined class centroids for image classification.
problem Reducing the need for labeled data in deep learning.
method Use a small number of labeled samples and data augmentation on unlabeled samples. Constrain all samples to predefined evenly-distributed class centroids (PEDCC) using loss functions.
result Achieves state-of-the-art results with minimal labeled data.
Signal recovery from unlabeled samples using a novel duality with Compressed Sensing.
problem Recovering a signal from unlabeled linear projections.
method Developed a duality between unlabeled sensing and Compressed Sensing, introduced a Restricted Isometry Property (RIP), and designed an Alternating Minimization algorithm.
result Signal recovery is possible with more samples than the signal dimension, similar to Compressed Sensing.
A new method for domain generalization using unlabeled data.
problem Learning from multiple domains with limited labeled data.
method Combines meta learning and semi-supervised learning with entropy-based pseudo-labeling and discrepancy loss.
result Significantly outperforms state-of-the-art methods on benchmark datasets.
New method uses unlabeled data to estimate intercept in case-control logistic regression.
problem Estimating intercept in case-control logistic regression.
method Construct likelihood function, use iterative algorithm to find maximum likelihood estimator.
result Proposed method identifies intercept and improves estimation efficiency.
MTL method uses unlabeled data with pseudo labels to improve classification with disjoint datasets.
problem Improving classification performance with disjoint labeled datasets using unlabeled data.
method Proposes MTL-SA method to select and augment unlabeled data with confident pseudo labels and close distribution to labeled data.
result Extensive experiments show the effectiveness of MTL-SA method in improving classification performance.
Active learning selects most informative unlabeled samples for labeling.
problem Efficiently label unlabeled data in applications with scarce labeled data.
method Formulated as open-set recognition, uses VNNs to identify uncertain samples.
result Achieved state-of-the-art results on MNIST, CIFAR-10, and CIFAR-100.
LACD uses unlabeled data to improve conditional diffusion models.
problem Costly and time-consuming acquisition of labeled data.
method Label-augmented conditional diffusion (LACD) with joint denoising score matching.
result LACD converges faster in total variation and Wasserstein-1 distances with sufficient unlabeled data.
A robust framework maximizes AUC with outlier detection and feature selection for PU classification.
problem Challenges in PU classification, especially with complex data and mislabeled/unlabeled samples.
method Unified AUC maximization, outlier detection, and feature selection.
result Generalization error bounds and practical guidance for model training.
Curriculum Labeling improves semi-supervised learning with pseudo-labeling, achieving high accuracy with minimal labeled data.
problem Improving semi-supervised learning with limited labeled data.
method Applying curriculum learning principles and restarting model parameters before each self-training cycle.
result 94.91% accuracy on CIFAR-10 with only 4,000 labeled samples.
New method corrects output-side domain shift without labeled data.
problem Learning mapping from input to output space with domain shift.
method Adapts learned function using unlabeled outputs from target distribution.
result Derives generalization bounds using discrepancy terms.
Study improves treatment effect estimation using unlabeled covariates.
problem Estimating treatment effects with limited labeled data.
method Developed efficiency bounds and estimators for semi-supervised setting.
result Estimators using unlabeled covariates have lower asymptotic variance.
A new method estimates population mean using labeled and unlabeled data.
problem Estimating population mean with limited labeled data.
method Semi-supervised inference framework, least squares method.
result Proposed estimators outperform ordinary sample mean.
Estimates class prior for unlabeled data using kernel embedding.
problem Estimating class prior in PU learning scenario where only positive and full population samples are available.
method Direct estimator based on distribution matching and kernel embedding in Reproducing Kernel Hilbert Space.
result Asymptotic consistency and explicit deviation bound for the estimator.
This paper proves transductive learning is as hard as inductive learning.
problem The difficulty of transductive classification compared to inductive learning.
method Proves minimax lower bounds for transductive, realizable, binary classification.
result Transductive learning is as hard as inductive learning, with a lower bound of Ω(d/ε+log(1/δ)/ε) for m≤u. Medix uses the median to detect outliers from unlabeled data for robust OOD detection.
problem Challenges in using unlabeled data for OOD detection due to mixed InD and OOD samples.
method Introduces Medix, a framework using the median operation to identify outliers from unlabeled data.
result Empirical results show Medix outperforms existing methods in open-world settings.
Assessing the performance of a learned model is a crucial part of machine learning. However, in some domains only positive and unlabeled examples are available, which prohibits the use of most standard evaluation metrics. We propose an approach to estimate any metric based on contingency tables, including ROC and PR cu…
Framework uses unlabeled out-of-domain data to improve semi-supervised classification.
problem Improving generalization in semi-supervised classification problems.
method Combines Distributionally Robust Optimization (DRO) with self-supervised training.
result Significant improvement in generalization error compared to ERM.
EAGC boosts GCD by regulating gradient entanglement, improving known and novel category separability.
problem Gradient entanglement distorts supervised gradients and overlaps known and novel class representations.
method EAGC uses AGA and EEP to align and project gradients, reducing entanglement and overlap.
result EAGC consistently boosts GCD performance, setting new state-of-the-art results.
ALIS uses probabilistic sampling to minimize true loss in active learning.
problem Efficiently choosing data points to label in active learning.
method Proposes ALIS algorithm with optimal sampling distribution.
result Derives upper bounds on true loss for any sampling procedure.
Research improves open-set learning by leveraging unlabelled data.
problem Learning between observed and unobserved novel categories.
method Unified policy of positive and unlabelled learning, semi-supervised learning, and open-set recognition.
result Achieves state-of-the-art results in open-set learning.
Study public-data assisted private stochastic optimization with labeled or unlabeled public data.
problem Limits and capability of public-data assisted differentially private (PA-DP) algorithms in stochastic convex optimization.
method Lower bounds for PA-DP mean estimation and novel methods for leveraging public data in private supervised learning.
result Achieved dimension independent rate for GLM with unlabeled public data, showing optimality.
GPNs use unlabeled data to estimate uncertainty in Bayesian problems.
problem Limited training data in high-dimensional problems.
method Generative Posterior Networks (GPNs) that approximate the posterior distribution using unlabeled data.
result GPNs improve epistemic uncertainty estimation and scalability.
Paper shows using unlabeled mismatched images can improve KD for image classification.
problem Improving Knowledge Distillation for image classification with limited labeled data.
method Used unlabeled mismatched images as stimulus for KD, focusing on stimulus complexity.
result Stimulus complexity is crucial for KD's effectiveness, as demonstrated on MNIST and CIFAR datasets.
Proposes a new semi-supervised learning method to reduce distribution mismatch.
problem Empirical distribution mismatch between labeled and unlabeled data in semi-supervised learning.
method Adversarial training and interpolation strategy to align labeled and unlabeled data distributions.
result Demonstrates improved performance on benchmark datasets SVHN and CIFAR10.
A new method for PU learning improves classification error on CIFAR-10.
problem Learning from positive and unlabeled data in practical applications.
method A simple yet effective data augmentation method based on consistency regularization.
result Achieves an averaged improvement of 3.40 points in classification error on CIFAR-10.
This paper uses unlabeled data to improve compressed neural networks.
problem Difficulty in retraining pre-trained models due to limited labeled data.
method Uses unlabeled data to mimic classification characteristics and aligns feature distributions using adversarial loss.
result Unlabeled data significantly improves the performance of compressed neural networks.
In-context learning solves PU classification without iterative optimization.
problem Binary classification with only labeled positives and unlabeled samples.
method Pretrained transformer (PUICL) that learns from synthetic PU datasets.
result Outperforms four standard PU learning baselines on 20 benchmarks.
New method uses unlabeled data to prevent forgetting in deep learning.
problem Catastrophic forgetting in lifelong learning with deep neural networks.
method Class-incremental learning scheme with global distillation, confidence-based sampling, and learning strategy.
result Significantly higher accuracy and less forgetting compared to state-of-the-art methods.
Generative Adversarial Positive-Unlabeled Learning tackles binary PU data classification.
problem Overfitting in conventional PU models with limited positive data.
method Innovative generative adversarial framework with multiple discriminators and generators.
result GenPU framework can recover both positive and negative data distributions at equilibrium.
ORDisCo learns from unlabeled data to improve semi-supervised continual learning.
problem Lack of effective use of unlabeled data in semi-supervised continual learning.
method Deep Online Replay with Discriminator Consistency (ORDisCo) that continually passes the learned data distribution to a classifier and selectively stabilizes discriminator parameters.
result Significant performance improvement on various semi-supervised learning benchmark datasets.
SelectNet selectively adds unlabelled data to balance imbalanced training data.
problem Training from imbalanced data in machine learning.
method SelectNet trains a deep neural network to add unlabelled data with predicted labels to the training dataset.
result SelectNet effectively balances imbalanced training data, improving classifier performance.
Method transfers knowledge between partially labeled domains to classify all samples.
problem Weakly supervised open-set domain adaptation between partially labeled domains.
method Collaborative Distribution Alignment (CDA) method for bilaterally knowledge transfer and outlier identification.
result Achieves state-of-the-art performance on Office benchmark and person reidentification.
Two algorithms reduce label complexity in machine learning using stratified sampling.
problem Reducing the number of true labels needed for machine learning evaluation.
method Proposes two algorithms that estimate strata properties and optimize evaluation accuracy.
result Demonstrates algorithms are rate optimal and reduce label complexity.
Active learning selects samples for labeling to build accurate models with minimal labeled data.
problem Costly acquisition of labeled data in supervised learning.
method Adaptive selection of unlabeled data samples for labeling.
result Efficient model building with minimal labeled data.
Two new ALR approaches based on GS reduce labeled samples needed for regression.
problem Need substantial labeled samples for regression models, but unlabeled samples are easy to collect.
method Proposes two new ALR approaches based on greedy sampling (GS) to select beneficial unlabeled samples.
result Extensive experiments on various datasets verified the effectiveness and robustness of the approaches.
A new method predicts true classes from positive and unlabeled data with additional labeled observations.
problem Predicting true classes from positive and unlabeled data with selection bias.
method Introduces augmented PU prediction, allowing feature-dependent labeling, and compares various empirical Bayes rules.
result The variational autoencoder-based method performs similarly or better than other methods and improves accuracy for unlabeled samples.
SAL framework uses unlabeled data to improve OOD detection.
problem Lack of clean OOD samples makes OOD detection challenging.
method SAL framework separates candidate outliers and trains an OOD classifier.
result SAL achieves state-of-the-art performance on benchmarks.
Unlabeled data improves adversarial robustness in machine learning.
problem Improving robustness of machine learning models against adversarial attacks.
method Semisupervised learning with self-training using unlabeled data.
result Semisupervised learning with unlabeled data significantly improves adversarial robustness.