Repeated self-distillation improves model performance significantly.
problem How much gain is possible by applying multiple steps of self-distillation?
method Investigated linear regression tasks, applied multiple steps of self-distillation, analyzed excess risk reduction.
result Multi-step self-distillation reduces excess risk by a factor as large as d, where d is the input dimension. Improved self-distillation reduces label noise and enhances model accuracy.
problem Label noise in multi-class classification.
method Label averaging and refined partial labels.
result Single-round self-distillation achieves comparable performance to multi-round distillation.
Self-distillation optimally improves model performance in spiked covariance models.
problem Improving model performance in spiked covariance models.
method Developed spectral shrinkage estimators and analyzed self-distillation.
result Self-distillation achieves optimal performance among spectral shrinkage estimators for spiked covariance matrices.
Knowledge distillation introduced in the deep learning context is a method to transfer knowledge from one architecture to another. In particular, when the architectures are identical, this is called self-distillation. The idea is to feed in predictions of the trained model as new target values for retraining (and itera…
Self-distillation improves model performance by increasing teacher diversity and smoothing predictions.
problem Improving model generalization and performance through self-distillation.
method Interpreting self-distillation as MAP estimation and proposing instance-specific label smoothing.
result Self-distillation enhances model performance by increasing teacher diversity and smoothing predictions.
Improved generalization with iterative self-distillation using weighted ground-truth targets.
problem Improving generalization accuracy in neural networks.
method Iterative kernel regression with weighted ground-truth targets and ℓ2 regularization. result Closed-form solution for optimal weighting parameter and efficient estimation.
Optimal self-distillation improves generative models' velocity risk and mode recovery.
problem Improving generative models' velocity risk and mode recovery.
method Proved optimal self-distillation for rectified flow via linear probing, derived mixing coefficient, and provided validation tuning.
result Optimal self-distillation improves velocity risk and mode recovery.
Self-distillation improves constrained language generation by aligning models with target distributions.
problem Sparse and uninformative reward signals in constrained generation settings.
method Iteratively refining the base model through self-distillation, incorporating learned twist functions and proposals.
result Substantial gains in generation quality through improved model alignment with target distributions.
Study shows optimal self-distillation improves model performance on noisy data.
problem Improving model performance on noisy Gaussian mixture data.
method Hyperparameter-tuned multi-stage self-distillation with a linear classifier, using replica method.
result Primary driver of SD's performance improvement is denoising through hard pseudo-labels.
Self-Distilled Disentanglement improves counterfactual predictions by separating variables.
problem Improving counterfactual predictions in the presence of confounders and unobserved variables.
method Self-Distilled Disentanglement framework based on information theory.
result Effective counterfactual inference in synthetic and real-world datasets.
Convolutional neural networks have been widely deployed in various application scenarios. In order to extend the applications' boundaries to some accuracy-crucial domains, researchers have been investigating approaches to boost accuracy through either deeper or wider network structures, which brings with them the expon…
Self-distillation improves model performance in noisy label settings.
problem Improving model accuracy in supervised learning with noisy labels.
method Analyzes self-distillation in two supervised learning problems with noisy labels, using theoretical and empirical approaches.
result Optimal self-distillation parameter is greater than 1 in high label noise regimes, outperforming traditional methods.
Study on ensemble, distillation, and self-distillation in deep learning models.
problem Improving test accuracy in deep learning models using ensemble and distillation methods.
method Formal study of ensemble and distillation, considering multi-view data structure.
result Proven that ensemble and distillation can improve test accuracy in deep learning models, and the superior performance can be distilled into a single model.
Two approaches extend knowledge distillation to Gaussian Processes, showing relationships to existing methods.
problem Applying knowledge distillation to Gaussian Processes for regression and classification.
method Data-centric and distribution-centric approaches to extend distillation to GPR and GPC.
result Distribution-centric approach for GPC approximately corresponds to data duplication and scaling.
A new method for training GNNs without a teacher model.
problem Training over-parameterized GNN models is difficult and inefficient.
method GNN Self-Distillation (GNN-SD) with NDR and ADR.
result Improves GNN performance with less training cost and better generalization.
Self-augmentation improves deep networks for few-shot learning with minimal training data.
problem Improving deep networks' generalization to unseen classes with limited training examples.
method Self-augmentation using self-mix and self-distillation techniques, combined with regional dropout and local representation learning.
result The method outperforms state-of-the-art few-shot learning methods on prevalent benchmarks.
Optimal SD improves ridge regression performance strictly and precisely.
problem Improving ridge regression performance through self-distillation.
method Analyzes unconstrained SD for ridge regression, deriving optimal mixing weight and asymptotic risk.
result Optimal SD strictly improves ridge regression performance, with exact risk equivalents derived.
Distillation is a method to transfer knowledge from one model to another and often achieves higher accuracy with the same capacity. In this paper, we aim to provide a theoretical understanding on what mainly helps with the distillation. Our answer is "early stopping". Assuming that the teacher network is overparameteri…
SAIL improves graph node representation learning by distilling knowledge between graphs.
problem Improving graph node representation learning with GNNs in unsupervised scenarios.
method SAIL framework with intra- and inter-graph knowledge distillation.
result SAIL consistently outperforms state-of-the-art baselines on various benchmark datasets.
One-step diffusion samplers reduce sampling time and computational costs.
problem Efficient sampling from complex distributions.
method One-step diffusion, self-distillation, deterministic flow.
result Achieves competitive sample quality with fewer evaluations.
Self-supervised learning, which learns by constructing artificial labels given only the input signals, has recently gained considerable attention for learning representations with unlabeled datasets, i.e., learning without any human-annotated supervision. In this paper, we show that such a technique can be used to sign…
A new method for distilling predictions from a teacher model to a student model without original training data.
problem Training without original labeled data.
method Prediction-only distillation scheme for linear and logistic regression.
result Prediction mixing can outperform both the teacher and pure-distilled models.
Self-PU combines self-training with PU learning for improved binary classification.
problem Binary classification from unlabeled data and few labeled positives.
method Integrates self-paced training, self-calibrated loss, and self-distillation.
result Significantly improved performance on Alzheimer's Disease classification.
Self-training in linear models shows a U-shaped test-risk curve due to signal forgetting and denoising.
problem Understanding the dynamics of iterative self-training in high-dimensional linear regression.
method Derivation of deterministic-equivalent recursions for prediction risk and effective noise, analysis of signal forgetting and denoising effects.
result An optimal early-stopping time is determined, and a U-shaped test-risk curve is observed.
Proposes LsrKD and MrKD to improve neural network training performance.
problem Improving neural network training performance, especially on deep networks.
method Extends Label Smoothing Regularization with Self-Knowledge Distillation, introducing LsrKD and MrKD.
result LsrKD and MrKD significantly improve training performance on deep neural networks.
Transformer-based method discovers objects from images without labels.
problem Discovering objects in images without labeled data.
method Graph-based approach using self-supervised transformer features and normalized graph-cut.
result Significantly boosts performance in unsupervised object discovery.