Dataset distillation aims to compress a dataset into a much smaller one so that a model trained on the distilled dataset achieves high accuracy. Current methods frame this as maximizing the distilled classification accuracy for a budget of K distilled images-per-class, where K is a positive integer. In this paper, we push the boundaries of dataset distillation, compressing the dataset into less than an image-per-class. It is important to realize that the meaningful quantity is not the number of distilled images-per-class but the number of distilled pixels-per-dataset. We therefore, propose Poster Dataset Distillation (PoDD), a new approach that distills the entire original dataset into a single poster. The poster approach motivates new technical solutions for creating training images and learnable labels. Our method can achieve comparable or better performance with less than an image-per-class compared to existing methods that use one image-per-class. Specifically, our method establishes a new state-of-the-art performance on CIFAR-10, CIFAR-100, and CUB200 using as little as 0.3 images-per-class.
Dataset distillation (DD) allows datasets to be distilled to fractions of their original size while preserving the rich distributional information so that models trained on the distilled datasets can achieve a comparable accuracy while saving significant computational loads. Recent research in this area has been focusing on improving the accuracy of models trained on distilled datasets. In this paper, we aim to explore a new perspective of DD. We study how to embed adversarial robustness in distilled datasets, so that models trained on these datasets maintain the high accuracy and meanwhile acquire better adversarial robustness. We propose a new method that achieves this goal by incorporating curvature regularization into the distillation process with much less computational overhead than standard adversarial training. Extensive empirical experiments suggest that our method not only outperforms standard adversarial training on both accuracy and robustness with less computation overhead but is also capable of generating robust distilled datasets that can withstand various adversarial attacks.
The extensive amounts of data required for training deep neural networks pose significant challenges on storage and transmission fronts. Dataset distillation has emerged as a promising technique to condense the information of massive datasets into a much smaller yet representative set of synthetic samples. However, traditional dataset distillation approaches often struggle to scale effectively with high-resolution images and more complex architectures due to the limitations in bi-level optimization. Recently, several works have proposed exploiting knowledge distillation with decoupled optimization schemes to scale up dataset distillation. Although these methods effectively address the scalability issue, they rely on extensive image augmentations requiring the storage of soft labels for augmented images. In this paper, we introduce Dataset Distillation using Diffusion Models (D3M) as a novel paradigm for dataset distillation, leveraging recent advancements in generative text-to-image foundation models. Our approach utilizes textual inversion, a technique for fine-tuning text-to-image generative models, to create concise and informative representations for large datasets. By employing these learned text prompts, we can efficiently store and infer new samples for introducing data variability within a fixed memory budget. We show the effectiveness of our method through extensive experiments across various computer vision benchmark datasets with different memory budgets.
Machine learning traditionally relies on increasingly larger datasets. Yet, such datasets pose major storage challenges and usually contain non-influential samples, which could be ignored during training without negatively impacting the training quality. In response, the idea of distilling a dataset into a condensed set of synthetic samples, i.e., a distilled dataset, emerged. One key aspect is the selected architecture, usually ConvNet, for linking the original and synthetic datasets. However, the final accuracy is lower if the employed model architecture differs from that used during distillation. Another challenge is the generation of high-resolution images (128x128 and higher). To address both challenges, this paper proposes Latent Dataset Distillation with Diffusion Models (LD3M) that combine diffusion in latent space with dataset distillation. Our novel diffusion process is tailored for this task and significantly improves the gradient flow for distillation. By adjusting the number of diffusion steps, LD3M also offers a convenient way of controlling the trade-off between distillation speed and dataset quality. Overall, LD3M consistently outperforms state-of-the-art methods by up to 4.8 p.p. and 4.2 p.p. for 1 and 10 images per class, respectively, and on several ImageNet subsets and high resolutions (128x128 and 256x256).
What does a neural network learn when training from a task-specific dataset? Synthesizing this knowledge is the central idea behind Dataset Distillation, which recent work has shown can be used to compress large datasets into a small set of input-label pairs ($\textit{prototypes}$) that capture essential aspects of the original dataset. In this paper, we make the key observation that existing methods distilling into explicit prototypes are very often suboptimal, incurring in unexpected storage cost from distilled labels. In response, we propose $\textit{Distributional Dataset Distillation}$ (D3), which encodes the data using minimal sufficient per-class statistics and paired with a decoder, we distill dataset into a compact distributional representation that is more memory-efficient compared to prototype-based methods. To scale up the process of learning these representations, we propose $\textit{Federated distillation}$, which decomposes the dataset into subsets, distills them in parallel using sub-task experts and then re-aggregates them. We thoroughly evaluate our algorithm on a three-dimensional metric and show that our method achieves state-of-the-art results on TinyImageNet and ImageNet-1K. Specifically, we outperform the prior art by $6.9\%$ on ImageNet-1K under the storage budget of 2 images per class.
Dataset distillation, a pragmatic approach in machine learning, aims to create a smaller synthetic dataset from a larger existing dataset. However, existing distillation methods primarily adopt a model-based paradigm, where the synthetic dataset inherits model-specific biases, limiting its generalizability to alternative models. In response to this constraint, we propose a novel methodology termed "model pool". This approach involves selecting models from a diverse model pool based on a specific probability distribution during the data distillation process. Additionally, we integrate our model pool with the established knowledge distillation approach and apply knowledge distillation to the test process of the distilled dataset. Our experimental results validate the effectiveness of the model pool approach across a range of existing models while testing, demonstrating superior performance compared to existing methodologies.
Dataset distillation (DD) has emerged as a widely adopted technique for crafting a synthetic dataset that captures the essential information of a training dataset, facilitating the training of accurate neural models. Its applications span various domains, including transfer learning, federated learning, and neural architecture search. The most popular methods for constructing the synthetic data rely on matching the convergence properties of training the model with the synthetic dataset and the training dataset. However, using the empirical loss as the criterion must be thought of as auxiliary in the same sense that the training set is an approximate substitute for the population distribution, and the latter is the data of interest. Yet despite its popularity, an aspect that remains unexplored is the relationship of DD to its generalization, particularly across uncommon subgroups. That is, how can we ensure that a model trained on the synthetic dataset performs well when faced with samples from regions with low population density? Here, the representativeness and coverage of the dataset become salient over the guaranteed training error at inference. Drawing inspiration from distributionally robust optimization, we introduce an algorithm that combines clustering with the minimization of a risk measure on the loss to conduct DD. We provide a theoretical rationale for our approach and demonstrate its effective generalization and robustness across subgroups through numerical experiments. The source code is available at https://github.com/Mming11/RobustDatasetDistillation.
Herein, we propose a novel dataset distillation method for constructing small informative datasets that preserve the information of the large original datasets. The development of deep learning models is enabled by the availability of large-scale datasets. Despite unprecedented success, large-scale datasets considerably increase the storage and transmission costs, resulting in a cumbersome model training process. Moreover, using raw data for training raises privacy and copyright concerns. To address these issues, a new task named dataset distillation has been introduced, aiming to synthesize a compact dataset that retains the essential information from the large original dataset. State-of-the-art (SOTA) dataset distillation methods have been proposed by matching gradients or network parameters obtained during training on real and synthetic datasets. The contribution of different network parameters to the distillation process varies, and uniformly treating them leads to degraded distillation performance. Based on this observation, we propose an importance-aware adaptive dataset distillation (IADD) method that can improve distillation performance by automatically assigning importance weights to different network parameters during distillation, thereby synthesizing more robust distilled datasets. IADD demonstrates superior performance over other SOTA dataset distillation methods based on parameter matching on multiple benchmark datasets and outperforms them in terms of cross-architecture generalization. In addition, the analysis of self-adaptive weights demonstrates the effectiveness of IADD. Furthermore, the effectiveness of IADD is validated in a real-world medical application such as COVID-19 detection.
Dataset distillation (DD) aims to synthesize a small dataset whose test performance is comparable to a full dataset using the same model. State-of-the-art (SoTA) methods optimize synthetic datasets primarily by matching heuristic indicators extracted from two networks: one from real data and one from synthetic data (see Fig.1, Left), such as gradients and training trajectories. DD is essentially a compression problem that emphasizes maximizing the preservation of information contained in the data. We argue that well-defined metrics which measure the amount of shared information between variables in information theory are necessary for success measurement but are never considered by previous works. Thus, we introduce mutual information (MI) as the metric to quantify the shared information between the synthetic and the real datasets, and devise MIM4DD numerically maximizing the MI via a newly designed optimizable objective within a contrastive learning framework to update the synthetic dataset. Specifically, we designate the samples in different datasets that share the same labels as positive pairs and vice versa negative pairs. Then we respectively pull and push those samples in positive and negative pairs into contrastive space via minimizing NCE loss. As a result, the targeted MI can be transformed into a lower bound represented by feature maps of samples, which is numerically feasible. Experiment results show that MIM4DD can be implemented as an add-on module to existing SoTA DD methods.
Dataset distillation is the technique of synthesizing smaller condensed datasets from large original datasets while retaining necessary information to persist the effect. In this paper, we approach the dataset distillation problem from a novel perspective: we regard minimizing the prediction discrepancy on the real data distribution between models, which are respectively trained on the large original dataset and on the small distilled dataset, as a conduit for condensing information from the raw data into the distilled version. An adversarial framework is proposed to solve the problem efficiently. In contrast to existing distillation methods involving nested optimization or long-range gradient unrolling, our approach hinges on single-level optimization. This ensures the memory efficiency of our method and provides a flexible tradeoff between time and memory budgets, allowing us to distil ImageNet-1K using a minimum of only 6.5GB of GPU memory. Under the optimal tradeoff strategy, it requires only 2.5$\times$ less memory and 5$\times$ less runtime compared to the state-of-the-art. Empirically, our method can produce synthetic datasets just 10% the size of the original, yet achieve, on average, 94% of the test accuracy of models trained on the full original datasets including ImageNet-1K, significantly surpassing state-of-the-art. Additionally, extensive tests reveal that our distilled datasets excel in cross-architecture generalization capabilities.
The poor cross-architecture generalization of dataset distillation greatly weakens its practical significance. This paper attempts to mitigate this issue through an empirical study, which suggests that the synthetic datasets undergo an inductive bias towards the distillation model. Therefore, the evaluation model is strictly confined to having similar architectures of the distillation model. We propose a novel method of EvaLuation with distillation Feature (ELF), which utilizes features from intermediate layers of the distillation model for the cross-architecture evaluation. In this manner, the evaluation model learns from bias-free knowledge therefore its architecture becomes unfettered while retaining performance. By performing extensive experiments, we successfully prove that ELF can well enhance the cross-architecture generalization of current DD methods. Code of this project is at \url{https://github.com/Lirui-Zhao/ELF}.
Contemporary machine learning requires training large neural networks on massive datasets and thus faces the challenges of high computational demands. Dataset distillation, as a recent emerging strategy, aims to compress real-world datasets for efficient training. However, this line of research currently struggle with large-scale and high-resolution datasets, hindering its practicality and feasibility. To this end, we re-examine the existing dataset distillation methods and identify three properties required for large-scale real-world applications, namely, realism, diversity, and efficiency. As a remedy, we propose RDED, a novel computationally-efficient yet effective data distillation paradigm, to enable both diversity and realism of the distilled data. Extensive empirical results over various neural architectures and datasets demonstrate the advancement of RDED: we can distill the full ImageNet-1K to a small dataset comprising 10 images per class within 7 minutes, achieving a notable 42% top-1 accuracy with ResNet-18 on a single RTX-4090 GPU (while the SOTA only achieves 21% but requires 6 hours).
Data heterogeneity presents significant challenges for federated learning (FL). Recently, dataset distillation techniques have been introduced, and performed at the client level, to attempt to mitigate some of these challenges. In this paper, we propose a highly efficient FL dataset distillation framework on the server side, significantly reducing both the computational and communication demands on local devices while enhancing the clients' privacy. Unlike previous strategies that perform dataset distillation on local devices and upload synthetic data to the server, our technique enables the server to leverage prior knowledge from pre-trained deep generative models to synthesize essential data representations from a heterogeneous model architecture. This process allows local devices to train smaller surrogate models while enabling the training of a larger global model on the server, effectively minimizing resource utilization. We substantiate our claim with a theoretical analysis, demonstrating the asymptotic resemblance of the process to the hypothetical ideal of completely centralized training on a heterogeneous dataset. Empirical evidence from our comprehensive experiments indicates our method's superiority, delivering an accuracy enhancement of up to 40% over non-dataset-distillation techniques in highly heterogeneous FL contexts, and surpassing existing dataset-distillation methods by 18%. In addition to the high accuracy, our framework converges faster than the baselines because rather than the server trains on several sets of heterogeneous data distributions, it trains on a multi-modal distribution. Our code is available at https://github.com/FedDG23/FedDG-main.git
Recently, dataset distillation has paved the way towards efficient machine learning, especially for image datasets. However, the distillation for videos, characterized by an exclusive temporal dimension, remains an underexplored domain. In this work, we provide the first systematic study of video distillation and introduce a taxonomy to categorize temporal compression. Our investigation reveals that the temporal information is usually not well learned during distillation, and the temporal dimension of synthetic data contributes little. The observations motivate our unified framework of disentangling the dynamic and static information in the videos. It first distills the videos into still images as static memory and then compensates the dynamic and motion information with a learnable dynamic memory block. Our method achieves state-of-the-art on video datasets at different scales, with a notably smaller memory storage budget. Our code is available at https://github.com/yuz1wan/video_distillation.
In many applications, Neural Nets (NNs) have classification performance on par or even exceeding human capacity. Moreover, it is likely that NNs leverage underlying features that might differ from those humans perceive to classify. Can we "reverse-engineer" pertinent features to enhance our scientific understanding? Here, we apply this idea to the notoriously difficult task of galaxy classification: NNs have reached high performance for this task, but what does a neural net (NN) "see" when it classifies galaxies? Are there morphological features that the human eye might overlook that could help with the task and provide new insights? Can we visualize tracers of early evolution, or additionally incorporated spectral data? We present a novel way to summarize and visualize galaxy morphology through the lens of neural networks, leveraging Dataset Distillation, a recent deep-learning methodology with the primary objective to distill knowledge from a large dataset and condense it into a compact synthetic dataset, such that a model trained on this synthetic dataset achieves performance comparable to a model trained on the full dataset. We curate a class-balanced, medium-size high-confidence version of the Galaxy Zoo 2 dataset, and proceed with dataset distillation from our accurate NN-classifier to create synthesized prototypical images of galaxy morphological features, demonstrating its effectiveness. Of independent interest, we introduce a self-adaptive version of the state-of-the-art Matching Trajectory algorithm to automate the distillation process, and show enhanced performance on computer vision benchmarks.
Dataset distillation or condensation aims to generate a smaller but representative subset from a large dataset, which allows a model to be trained more efficiently, meanwhile evaluating on the original testing data distribution to achieve decent performance. Previous decoupled methods like SRe$^2$L simply use a unified gradient update scheme for synthesizing data from Gaussian noise, while, we notice that the initial several update iterations will determine the final outline of synthesis, thus an improper gradient update strategy may dramatically affect the final generation quality. To address this, we introduce a simple yet effective global-to-local gradient refinement approach enabled by curriculum data augmentation ($\texttt{CDA}$) during data synthesis. The proposed framework achieves the current published highest accuracy on both large-scale ImageNet-1K and 21K with 63.2% under IPC (Images Per Class) 50 and 36.1% under IPC 20, using a regular input resolution of 224$\times$224 with faster convergence speed and less synthetic time. The proposed model outperforms the current state-of-the-art methods like SRe$^2$L, TESLA, and MTT by more than 4% Top-1 accuracy on ImageNet-1K/21K and for the first time, reduces the gap to its full-data training counterparts to less than absolute 15%. Moreover, this work represents the inaugural success in dataset distillation on the larger-scale ImageNet-21K dataset under the standard 224$\times$224 resolution. Our code and distilled ImageNet-21K dataset of 20 IPC, 2K recovery budget are available at https://github.com/VILA-Lab/SRe2L/tree/main/CDA.
Dataset Distillation (DD) emerges as a powerful strategy to encapsulate the expansive information of large datasets into significantly smaller, synthetic equivalents, thereby preserving model performance with reduced computational overhead. Pursuing this objective, we introduce the Wasserstein distance, a metric grounded in optimal transport theory, to enhance distribution matching in DD. Our approach employs the Wasserstein barycenter to provide a geometrically meaningful method for quantifying distribution differences and capturing the centroid of distribution sets efficiently. By embedding synthetic data in the feature spaces of pretrained classification models, we facilitate effective distribution matching that leverages prior knowledge inherent in these models. Our method not only maintains the computational advantages of distribution matching-based techniques but also achieves new state-of-the-art performance across a range of high-resolution datasets. Extensive testing demonstrates the effectiveness and adaptability of our method, underscoring the untapped potential of Wasserstein metrics in dataset distillation.
This paper presents FreD, a novel parameterization method for dataset distillation, which utilizes the frequency domain to distill a small-sized synthetic dataset from a large-sized original dataset. Unlike conventional approaches that focus on the spatial domain, FreD employs frequency-based transforms to optimize the frequency representations of each data instance. By leveraging the concentration of spatial domain information on specific frequency components, FreD intelligently selects a subset of frequency dimensions for optimization, leading to a significant reduction in the required budget for synthesizing an instance. Through the selection of frequency dimensions based on the explained variance, FreD demonstrates both theoretical and empirical evidence of its ability to operate efficiently within a limited budget, while better preserving the information of the original dataset compared to conventional parameterization methods. Furthermore, based on the orthogonal compatibility of FreD with existing methods, we confirm that FreD consistently improves the performances of existing distillation methods over the evaluation scenarios with different benchmark datasets. We release the code at https://github.com/sdh0818/FreD.
Training on large-scale graphs has achieved remarkable results in graph representation learning, but its cost and storage have attracted increasing concerns. Existing graph condensation methods primarily focus on optimizing the feature matrices of condensed graphs while overlooking the impact of the structure information from the original graphs. To investigate the impact of the structure information, we conduct analysis from the spectral domain and empirically identify substantial Laplacian Energy Distribution (LED) shifts in previous works. Such shifts lead to poor performance in cross-architecture generalization and specific tasks, including anomaly detection and link prediction. In this paper, we propose a novel Structure-broadcasting Graph Dataset Distillation (SGDD) scheme for broadcasting the original structure information to the generation of the synthetic one, which explicitly prevents overlooking the original structure information. Theoretically, the synthetic graphs by SGDD are expected to have smaller LED shifts than previous works, leading to superior performance in both cross-architecture settings and specific tasks. We validate the proposed SGDD across 9 datasets and achieve state-of-the-art results on all of them: for example, on the YelpChi dataset, our approach maintains 98.6% test accuracy of training on the original graph dataset with 1,000 times saving on the scale of the graph. Moreover, we empirically evaluate there exist 17.6% ~ 31.4% reductions in LED shift crossing 9 datasets. Extensive experiments and analysis verify the effectiveness and necessity of the proposed designs. The code is available in the GitHub repository: https://github.com/RingBDStack/SGDD.
We present a new dataset condensation framework termed Squeeze, Recover and Relabel (SRe$^2$L) that decouples the bilevel optimization of model and synthetic data during training, to handle varying scales of datasets, model architectures and image resolutions for efficient dataset condensation. The proposed method demonstrates flexibility across diverse dataset scales and exhibits multiple advantages in terms of arbitrary resolutions of synthesized images, low training cost and memory consumption with high-resolution synthesis, and the ability to scale up to arbitrary evaluation network architectures. Extensive experiments are conducted on Tiny-ImageNet and full ImageNet-1K datasets. Under 50 IPC, our approach achieves the highest 42.5% and 60.8% validation accuracy on Tiny-ImageNet and ImageNet-1K, outperforming all previous state-of-the-art methods by margins of 14.5% and 32.9%, respectively. Our approach also surpasses MTT in terms of speed by approximately 52$\times$ (ConvNet-4) and 16$\times$ (ResNet-18) faster with less memory consumption of 11.6$\times$ and 6.4$\times$ during data synthesis. Our code and condensed datasets of 50, 200 IPC with 4K recovery budget are available at https://github.com/VILA-Lab/SRe2L.