Can pre-trained models assist in dataset distillation?

Dataset Distillation (DD) is a prominent technique that encapsulates knowledge from a large-scale original dataset into a small synthetic dataset for efficient training. Meanwhile, Pre-trained Models (PTMs) function as knowledge repositories, containing extensive information from the original dataset. This naturally raises a question: Can PTMs effectively transfer knowledge to synthetic datasets, guiding DD accurately? To this end, we conduct preliminary experiments, confirming the contribution of PTMs to DD. Afterwards, we systematically study different options in PTMs, including initialization parameters, model architecture, training epoch and domain knowledge, revealing that: 1) Increasing model diversity enhances the performance of synthetic datasets; 2) Sub-optimal models can also assist in DD and outperform well-trained ones in certain cases; 3) Domain-specific PTMs are not mandatory for DD, but a reasonable domain match is crucial. Finally, by selecting optimal options, we significantly improve the cross-architecture generalization over baseline DD methods. We hope our work will facilitate researchers to develop better DD techniques. Our code is available at https://github.com/yaolu-zjut/DDInterpreter.

Boosting Unsupervised Contrastive Learning Using Diffusion-Based Data Augmentation From Scratch

Unsupervised contrastive learning methods have recently seen significant improvements, particularly through data augmentation strategies that aim to produce robust and generalizable representations. However, prevailing data augmentation methods, whether hand designed or based on foundation models, tend to rely heavily on prior knowledge or external data. This dependence often compromises their effectiveness and efficiency. Furthermore, the applicability of most existing data augmentation strategies is limited when transitioning to other research domains, especially science-related data. This limitation stems from the paucity of prior knowledge and labeled data available in these domains. To address these challenges, we introduce DiffAug-a novel and efficient Diffusion-based data Augmentation technique. DiffAug aims to ensure that the augmented and original data share a smoothed latent space, which is achieved through diffusion steps. Uniquely, unlike traditional methods, DiffAug first mines sufficient prior semantic knowledge about the neighborhood. This provides a constraint to guide the diffusion steps, eliminating the need for labels, external data/models, or prior knowledge. Designed as an architecture-agnostic framework, DiffAug provides consistent improvements. Specifically, it improves image classification and clustering accuracy by 1.6%~4.5%. When applied to biological data, DiffAug improves performance by up to 10.1%, with an average improvement of 5.8%. DiffAug shows good performance in both vision and biological domains.

Color Prompting for Data-Free Continual Unsupervised Domain Adaptive Person Re-Identification

Unsupervised domain adaptive person re-identification (Re-ID) methods alleviate the burden of data annotation through generating pseudo supervision messages. However, real-world Re-ID systems, with continuously accumulating data streams, simultaneously demand more robust adaptation and anti-forgetting capabilities. Methods based on image rehearsal addresses the forgetting issue with limited extra storage but carry the risk of privacy leakage. In this work, we propose a Color Prompting (CoP) method for data-free continual unsupervised domain adaptive person Re-ID. Specifically, we employ a light-weighted prompter network to fit the color distribution of the current task together with Re-ID training. Then for the incoming new tasks, the learned color distribution serves as color style transfer guidance to transfer the images into past styles. CoP achieves accurate color style recovery for past tasks with adequate data diversity, leading to superior anti-forgetting effects compared with image rehearsal methods. Moreover, CoP demonstrates strong generalization performance for fast adaptation into new domains, given only a small amount of unlabeled images. Extensive experiments demonstrate that after the continual training pipeline the proposed CoP achieves 6.7% and 8.1% average rank-1 improvements over the replay method on seen and unseen domains, respectively. The source code for this work is publicly available in https://github.com/vimar-gu/ColorPromptReID.

Dataset Quantization

State-of-the-art deep neural networks are trained with large amounts (millions or even billions) of data. The expensive computation and memory costs make it difficult to train them on limited hardware resources, especially for recent popular large language models (LLM) and computer vision models (CV). Recent popular dataset distillation methods are thus developed, aiming to reduce the number of training samples via synthesizing small-scale datasets via gradient matching. However, as the gradient calculation is coupled with the specific network architecture, the synthesized dataset is biased and performs poorly when used for training unseen architectures. To address these limitations, we present dataset quantization (DQ), a new framework to compress large-scale datasets into small subsets which can be used for training any neural network architectures. Extensive experiments demonstrate that DQ is able to generate condensed small datasets for training unseen network architectures with state-of-the-art compression ratios for lossless model training. To the best of our knowledge, DQ is the first method that can successfully distill large-scale datasets such as ImageNet-1k with a state-of-the-art compression ratio. Notably, with 60% data from ImageNet and 20% data from Alpaca's instruction tuning data, the models can be trained with negligible or no performance drop for both vision tasks (including classification, semantic segmentation, and object detection) as well as language tasks (including instruction tuning tasks such as BBH and DROP).

The Snowflake Hypothesis: Training Deep GNN with One Node One Receptive field

Despite Graph Neural Networks demonstrating considerable promise in graph representation learning tasks, GNNs predominantly face significant issues with over-fitting and over-smoothing as they go deeper as models of computer vision realm. In this work, we conduct a systematic study of deeper GNN research trajectories. Our findings indicate that the current success of deep GNNs primarily stems from (I) the adoption of innovations from CNNs, such as residual/skip connections, or (II) the tailor-made aggregation algorithms like DropEdge. However, these algorithms often lack intrinsic interpretability and indiscriminately treat all nodes within a given layer in a similar manner, thereby failing to capture the nuanced differences among various nodes. To this end, we introduce the Snowflake Hypothesis -- a novel paradigm underpinning the concept of ``one node, one receptive field''. The hypothesis draws inspiration from the unique and individualistic patterns of each snowflake, proposing a corresponding uniqueness in the receptive fields of nodes in the GNNs. We employ the simplest gradient and node-level cosine distance as guiding principles to regulate the aggregation depth for each node, and conduct comprehensive experiments including: (1) different training schemes; (2) various shallow and deep GNN backbones, and (3) various numbers of layers (8, 16, 32, 64) on multiple benchmarks (six graphs including dense graphs with millions of nodes); (4) compare with different aggregation strategies. The observational results demonstrate that our hypothesis can serve as a universal operator for a range of tasks, and it displays tremendous potential on deep GNNs. It can be applied to various GNN frameworks, enhancing its effectiveness when operating in-depth, and guiding the selection of the optimal network depth in an explainable and generalizable way.

Learning Referring Video Object Segmentation from Weak Annotation

Referring video object segmentation (RVOS) is a task that aims to segment the target object in all video frames based on a sentence describing the object. Previous RVOS methods have achieved significant performance with densely-annotated datasets, whose construction is expensive and time-consuming. To relieve the burden of data annotation while maintaining sufficient supervision for segmentation, we propose a new annotation scheme, in which we label the frame where the object first appears with a mask and use bounding boxes for the subsequent frames. Based on this scheme, we propose a method to learn from this weak annotation. Specifically, we design a cross frame segmentation method, which uses the language-guided dynamic filters to thoroughly leverage the valuable mask annotation and bounding boxes. We further develop a bi-level contrastive learning method to encourage the model to learn discriminative representation at the pixel level. Extensive experiments and ablative analyses show that our method is able to achieve competitive performance without the demand of dense mask annotation. The code will be available at https://github.com/wangbo-zhao/WRVOS/.

CAME: Confidence-guided Adaptive Memory Efficient Optimization

Adaptive gradient methods, such as Adam and LAMB, have demonstrated excellent performance in the training of large language models. Nevertheless, the need for adaptivity requires maintaining second-moment estimates of the per-parameter gradients, which entails a high cost of extra memory overheads. To solve this problem, several memory-efficient optimizers (e.g., Adafactor) have been proposed to obtain a drastic reduction in auxiliary memory usage, but with a performance penalty. In this paper, we first study a confidence-guided strategy to reduce the instability of existing memory efficient optimizers. Based on this strategy, we propose CAME to simultaneously achieve two goals: fast convergence as in traditional adaptive methods, and low memory usage as in memory-efficient methods. Extensive experiments demonstrate the training stability and superior performance of CAME across various NLP tasks such as BERT and GPT-2 training. Notably, for BERT pre-training on the large batch size of 32,768, our proposed optimizer attains faster convergence and higher accuracy compared with the Adam optimizer. The implementation of CAME is publicly available.

Summarizing Stream Data for Memory-Restricted Online Continual Learning

Replay-based methods have proved their effectiveness on online continual learning by rehearsing past samples from an auxiliary memory. With many efforts made on improving training schemes based on the memory, however, the information carried by each sample in the memory remains under-investigated. Under circumstances with restricted storage space, the informativeness of the memory becomes critical for effective replay. Although some works design specific strategies to select representative samples, by only employing original images, the storage space is still not well utilized. To this end, we propose to Summarize the knowledge from the Stream Data (SSD) into more informative samples by distilling the training characteristics of real images. Through maintaining the consistency of training gradients and relationship to the past tasks, the summarized samples are more representative for the stream data compared to the original images. Extensive experiments are conducted on multiple online continual learning benchmarks to support that the proposed SSD method significantly enhances the replay effects. We demonstrate that with limited extra computational overhead, SSD provides more than 3% accuracy boost for sequential CIFAR-100 under extremely restricted memory buffer. The code is available in https://github.com/vimar-gu/SSD.

To Repeat or Not To Repeat: Insights from Scaling LLM under Token-Crisis

Recent research has highlighted the importance of dataset size in scaling language models. However, large language models (LLMs) are notoriously token-hungry during pre-training, and high-quality text data on the web is approaching its scaling limit for LLMs. To further enhance LLMs, a straightforward approach is to repeat the pre-training data for additional epochs. In this study, we empirically investigate three key aspects under this approach. First, we explore the consequences of repeating pre-training data, revealing that the model is susceptible to overfitting, leading to multi-epoch degradation. Second, we examine the key factors contributing to multi-epoch degradation, finding that significant factors include dataset size, model parameters, and training objectives, while less influential factors consist of dataset quality and model FLOPs. Finally, we explore whether widely used regularization can alleviate multi-epoch degradation. Most regularization techniques do not yield significant improvements, except for dropout, which demonstrates remarkable effectiveness but requires careful tuning when scaling up the model size. Additionally, we discover that leveraging mixture-of-experts (MoE) enables cost-effective and efficient hyper-parameter tuning for computationally intensive dense LLMs with comparable trainable parameters, potentially impacting efficient LLM development on a broader scale.

Response Length Perception and Sequence Scheduling: An LLM-Empowered LLM Inference Pipeline

Large language models (LLMs) have revolutionized the field of AI, demonstrating unprecedented capacity across various tasks. However, the inference process for LLMs comes with significant computational costs. In this paper, we propose an efficient LLM inference pipeline that harnesses the power of LLMs. Our approach begins by tapping into the potential of LLMs to accurately perceive and predict the response length with minimal overhead. By leveraging this information, we introduce an efficient sequence scheduling technique that groups queries with similar response lengths into micro-batches. We evaluate our approach on real-world instruction datasets using the LLaMA-based model, and our results demonstrate an impressive 86% improvement in inference throughput without compromising effectiveness. Notably, our method is orthogonal to other inference acceleration techniques, making it a valuable addition to many existing toolkits (e.g., FlashAttention, Quantization) for LLM inference.