New methods distill data for deep networks efficiently.
problem Reduction of training data cost and inconvenience.
method Generative teaching networks, gradient matching, Implicit Function Theorem.
result New methods are computationally more efficient and improve model performance.
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.
SiD distills pretrained diffusion models into a fast one-step generator.
problem Efficiently distilling pretrained diffusion models into a fast generator.
method Reformulates forward diffusion processes as semi-implicit distributions and uses three score-related identities to create a loss mechanism.
result Achieves high FID performance and significantly reduces generation time.
The paper explains knowledge distillation by analyzing visual concepts in DNNs.
problem Understanding how knowledge distillation affects the learning of visual concepts in deep neural networks.
method The paper proposes three hypotheses and designs mathematical metrics to evaluate feature representations of DNNs.
result The hypotheses were verified through experiments on various DNNs.
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.
This study improves knowledge distillation for RNN-T models with noisy labels.
problem Challenges in distilling knowledge from RNN-T models with variable quality teachers.
method Full-sum distillation and sequence-level knowledge distillation.
result Full-sum distillation outperforms other methods for RNN-T models, especially for bad teachers.
A new method distills datasets more efficiently and effectively.
problem Achieving competitive performance on test data with a small synthetic dataset.
method Tackles dataset distillation as a bilevel optimization problem, introduces RaT-BPTT to stabilize gradients and speed up optimization.
result Establishes new state-of-the-art performance across various benchmarks.
TL-ANDI distills context from source data to improve transfer learning for TFMs.
problem Limited transfer learning due to context-size constraints and distribution shifts.
method TL-ANDI uses posterior-aware distillation to construct a compact source context and locally distills labels.
result Improves transfer performance by addressing context-size and distribution shifts.
RealUID distills matching models using real data without GANs.
problem Slow inference in matching models like diffusion and flow.
method RealUID is a universal distillation framework that incorporates real data into the distillation procedure without using GANs.
result RealUID offers a simple theoretical foundation that covers previous distillation methods for Flow Matching and Diffusion models.
Distillation speeds up classifier training and provides insights into its success.
problem Empirical success of knowledge distillation without theoretical explanation.
method Study of linear and deep linear classifiers, proving a generalization bound.
result Three key factors for distillation success: data geometry, optimization bias, strong monotonicity.
Compressed Federated Distillation reduces communication in federated learning.
problem Communication constraints in Federated Learning.
method Compressed Federated Distillation (CFD) leverages soft labels and quantization techniques.
result Reduces communication by more than 4 orders of magnitude compared to Federated Averaging.
Model distillation aims to distill the knowledge of a complex model into a simpler one. In this paper, we consider an alternative formulation called dataset distillation: we keep the model fixed and instead attempt to distill the knowledge from a large training dataset into a small one. The idea is to synthesize a smal…
FAST-DAD distills complex ensemble models into faster, more accurate individual models.
problem Deploying complex AutoML ensemble predictors on tabular data is slow, large, and opaque.
method Data augmentation strategy based on Gibbs sampling from a self-attention pseudolikelihood estimator.
result FAST-DAD distillation produces significantly better individual models than standard training.
Dataset distillation is a method for reducing dataset sizes by learning a small number of synthetic samples containing all the information of a large dataset. This has several benefits like speeding up model training, reducing energy consumption, and reducing required storage space. Currently, each synthetic sample is …
Efficiently distills pretrained text-to-image models without real data, improving FID and CLIP scores.
problem Slow iterative refinement process of diffusion-based text-to-image models.
method Guided Score identity Distillation with Long and Short Classifier-Free Guidance.
result Achieves state-of-the-art FID performance with competitive CLIP score.
Distillation (Hinton et al., 2015) and privileged information (Vapnik & Izmailov, 2015) are two techniques that enable machines to learn from other machines. This paper unifies these two techniques into generalized distillation, a framework to learn from multiple machines and data representations. We provide theoretica…
Optimal quantization improves dataset distillation for faster training.
problem Efficiently train models with synthetic data.
method Reformulate disentangled methods as optimal quantization problems.
result Better performance and generalization on ImageNet-1K and subsets.
Knowledge distillation improves model accuracy by mimicking teacher model probabilities.
problem Improving model accuracy through model compression.
method Casting knowledge distillation as a semiparametric inference problem, deriving new guarantees, and developing enhancements.
result Enhancements improve student performance by mitigating teacher overfitting and underfitting.
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. New method distills cloud models into edge-friendly ones.
problem Cloud-to-edge model compression with limited data exchange.
method Two-step workflow of deprivatization and distillation.
result Outperforms previous state-of-the-art approaches on various benchmarks.
This paper investigates teacher hacking during language model distillation and proposes methods to mitigate it.
problem Teacher hacking during language model distillation, leading to suboptimal performance.
method A controlled experimental setup involving an oracle LM, teacher LM, and student LM, using fixed offline or online data generation techniques.
result Data diversity is the key factor in preventing teacher hacking during distillation.
Recent variants improve knowledge distillation performance.
problem Improving the performance of knowledge distillation.
method Introducing additional components or changing the learning process.
result These variants have shown promising results.
Knowledge distillation (KD) is a technique to derive optimal performance from a small student network (SN) by distilling knowledge of a large teacher network (TN) and transferring the distilled knowledge to the small SN. Since a role of convolutional neural network (CNN) in KD is to embed a dataset so as to perform a g…
Black-box risk scoring models permeate our lives, yet are typically proprietary or opaque. We propose Distill-and-Compare, a model distillation and comparison approach to audit such models. To gain insight into black-box models, we treat them as teachers, training transparent student models to mimic the risk scores ass…
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.
A new distillation method transfers channel information from teacher to student.
problem Transfer knowledge from teacher to student with fewer parameters and calculations.
method Channel Distillation (CD) and Guided Knowledge Distillation (GKD) with loss decay.
result Achieved 27.68% top-1 error on ImageNet with ResNet18, outperforming state-of-the-art methods.
Paper analyzes dataset distillation for efficient encoding of task-relevant information.
problem Efficiently encoding task-relevant information from gradient-based learning of non-linear tasks.
method Theoretical analysis of dataset distillation applied to two-layer neural networks with gradient-based training.
result Low-dimensional structure of the problem is efficiently encoded into distilled data, reproducing a model with high generalization ability.
Knowledge distillation deals with the problem of training a smaller model (Student) from a high capacity source model (Teacher) so as to retain most of its performance. Existing approaches use either the training data or meta-data extracted from it in order to train the Student. However, accessing the dataset on which …
FED distills ensemble predictions efficiently, improving accuracy and uncertainty.
problem Efficiently distilling ensemble predictions from limited data.
method Proposes Functional Ensemble Distillation (FED) using mixup augmentation.
result Significantly boosts performance in accuracy and uncertainty estimation.
Over the past year, the emergence of transfer learning with large-scale language models (LM) has led to dramatic performance improvements across a broad range of natural language understanding tasks. However, the size and memory footprint of these large LMs makes them difficult to deploy in many scenarios (e.g. on mobi…
A new method for privacy-preserving data distillation using wavelet features from ScatterNet.
problem Privacy-preserving data distillation for small datasets.
method Differentially private kernel inducing points using wavelet features from ScatterNet.
result Achieves privacy-preserving data distillation with manageable computational cost.
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…
Paper introduces a simulator-free approach to reinforcement learning policy distillation.
problem Learning multiplicity of cases corresponding to a given action in reinforcement learning.
method Generative adversarial approach to find multiple exemplars for each output class.
result Improves over state-of-the-art on data-free learning of student networks.
Model compression is eminently suited for deploying deep learning on IoT-devices. However, existing model compression techniques rely on access to the original or some alternate dataset. In this paper, we address the model compression problem when no real data is available, e.g., when data is private. To this end, we p…
Consistency distillation reduces memorization in diffusion models without harming sample quality.
problem Understanding how distillation affects memorization in diffusion models.
method Analysis of consistency distillation in diffusion models using a random feature neural network model.
result Consistency distillation reduces memorization in diffusion models without harming sample quality.
New method for decentralized learning from diverse clients.
problem Achieving high-performance global model from diverse client models.
method Decentralized Learning via Adaptive Distillation (DLAD).
result Demonstrates effectiveness of DLAD on multiple public datasets.
Paper introduces CoD for few-shot task-aware knowledge distillation using counterfactual explanations.
problem Lack of data for task-aware distillation in resource-constrained scenarios.
method Counterfactual-explanation-infused Distillation CoD for few-shot task-aware knowledge distillation.
result CoD achieves superior performance with significantly fewer samples than baseline methods.
The paper analyzes knowledge distillation in wide neural networks, providing theoretical insights and practical implications.
problem Lack of theoretical understanding of knowledge distillation in wide neural networks.
method Theoretical analysis of knowledge distillation in a linearized model of a wide neural network, introducing a metric of task training difficulty.
result For a perfect teacher, a high ratio of teacher's soft labels can be beneficial. For imperfect teacher, hard labels can correct wrong predictions.
Current approaches for Knowledge Distillation (KD) either directly use training data or sample from the training data distribution. In this paper, we demonstrate effectiveness of 'mismatched' unlabeled stimulus to perform KD for image classification networks. For illustration, we consider scenarios where this is a comp…
This work proposes splitting deep neural networks into smaller sub-networks for faster and more efficient distillation.
problem Challenges in training deep neural networks, including local optima, gradient issues, and computational demands.
method Proposes a non-end-to-end distillation approach by splitting networks into smaller, independent sub-networks (neighbourhoods).
result Independent training of smaller sub-networks can speed up distillation and improve efficiency in various applications.
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.
Often we wish to transfer representational knowledge from one neural network to another. Examples include distilling a large network into a smaller one, transferring knowledge from one sensory modality to a second, or ensembling a collection of models into a single estimator. Knowledge distillation, the standard approa…
EM Distillation simplifies diffusion models to one-step generators.
problem Efficient sampling from complex diffusion models with minimal loss of quality.
method EM Distillation, a maximum likelihood approach based on Expectation-Maximization.
result EM Distillation outperforms existing one-step generative methods in FID scores.
DOSFL reduces federated learning communication by one round, preserving model performance.
problem High communication costs in federated learning with poorly distributed data.
method Clients distill their private data into synthetic data, sending only this to the server for training.
result Total communication cost up to 3 orders of magnitude less than FedAvg while maintaining comparable performance.
Ensembles of neural networks have been shown to give better performance than single networks, both in terms of predictions and uncertainty estimation. Additionally, ensembles allow the uncertainty to be decomposed into aleatoric (data) and epistemic (model) components, giving a more complete picture of the predictive u…
New federated learning protocols improve on knowledge distillation's poor performance.
problem Designing a universal API for federated learning without public data.
method Proposed Federated Kernel ridge regression using knowledge distillation.
result Performance of new protocols closely matches theoretical predictions.
Few-step distillation improves T2I models without real images or CFG trade-offs.
problem Challenges in accelerating T2I models with high-resolution and CFG.
method Score identity distillation (SiD) for few-step generation, with adversarial loss and new guidance strategies.
result State-of-the-art performance on SDXL at 1024x1024 resolution, robust to real images absence.
Survey of knowledge distillation for model compression.
problem Deploying large deep models on resource-limited devices.
method Knowledge distillation from large teacher models to small student models.
result Effective model compression and acceleration.