Soft-Label Dataset Distillation and Text Dataset Distillation
This work addresses the problem of reducing dataset sizes for faster and more efficient model training in machine learning, with incremental improvements in accuracy and new applications to text data.
The paper tackles dataset distillation by introducing soft labels to assign distributions instead of hard labels, improving accuracy by 2-4% on image tasks and enabling fewer samples than classes, such as achieving over 96% accuracy on MNIST with 10 distilled images. It also extends distillation to text datasets, showing models can retain near-original accuracy with just 20 distilled sentences on IMDB sentiment analysis.
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 assigned a single `hard' label, and also, dataset distillation can currently only be used with image data. We propose to simultaneously distill both images and their labels, thus assigning each synthetic sample a `soft' label (a distribution of labels). Our algorithm increases accuracy by 2-4% over the original algorithm for several image classification tasks. Using `soft' labels also enables distilled datasets to consist of fewer samples than there are classes as each sample can encode information for multiple classes. For example, training a LeNet model with 10 distilled images (one per class) results in over 96% accuracy on MNIST, and almost 92% accuracy when trained on just 5 distilled images. We also extend the dataset distillation algorithm to distill sequential datasets including texts. We demonstrate that text distillation outperforms other methods across multiple datasets. For example, models attain almost their original accuracy on the IMDB sentiment analysis task using just 20 distilled sentences. Our code can be found at $\href{https://github.com/ilia10000/dataset-distillation}{\text{https://github.com/ilia10000/dataset-distillation}}$.