Shaojie Li

Yifeng Shi, Feng Lv, Xinliang Wang et al.

With the continuous improvement of computing power and deep learning algorithms in recent years, the foundation model has grown in popularity. Because of its powerful capabilities and excellent performance, this technology is being adopted and applied by an increasing number of industries. In the intelligent transportation industry, artificial intelligence faces the following typical challenges: few shots, poor generalization, and a lack of multi-modal techniques. Foundation model technology can significantly alleviate the aforementioned issues. To address these, we designed the 1st Foundation Model Challenge, with the goal of increasing the popularity of foundation model technology in traffic scenarios and promoting the rapid development of the intelligent transportation industry. The challenge is divided into two tracks: all-in-one and cross-modal image retrieval. Furthermore, we provide a new baseline and benchmark for the two tracks, called Open-TransMind. According to our knowledge, Open-TransMind is the first open-source transportation foundation model with multi-task and multi-modal capabilities. Simultaneously, Open-TransMind can achieve state-of-the-art performance on detection, classification, and segmentation datasets of traffic scenarios. Our source code is available at https://github.com/Traffic-X/Open-TransMind.

Yuanfeng Wu, Shaojie Li, Zhiqiang Du et al.

Whole slide image (WSI) processing is becoming part of the key components of standard clinical diagnosis for various diseases. However, the direct application of conventional image processing algorithms to WSI faces certain obstacles because of WSIs' distinct property: the super-high resolution. The performance of most WSI-related tasks relies on the efficacy of the backbone which extracts WSI patch feature representations. Hence, we proposed BROW, a foundation model for extracting better feature representations for WSIs, which can be conveniently adapted to downstream tasks without or with slight fine-tuning. The model takes transformer architecture, pretrained using self-distillation framework. To improve model's robustness, techniques such as patch shuffling have been employed. Additionally, the model leverages the unique properties of WSIs, utilizing WSI's multi-scale pyramid to incorporate an additional global view, thereby further enhancing its performance. We used both private and public data to make up a large pretraining dataset, containing more than 11000 slides, over 180M extracted patches, encompassing WSIs related to various organs and tissues. To assess the effectiveness of \ourmodel, we run a wide range of downstream tasks, including slide-level subtyping, patch-level classification and nuclei instance segmentation. The results confirmed the efficacy, robustness and good generalization ability of the proposed model. This substantiates its potential as foundation model for WSI feature extraction and highlights promising prospects for its application in WSI processing.

Shaojie Li, Yong Liu

Gradient clipping is a commonly used technique to stabilize the training process of neural networks. A growing body of studies has shown that gradient clipping is a promising technique for dealing with the heavy-tailed behavior that emerged in stochastic optimization as well. While gradient clipping is significant, its theoretical guarantees are scarce. Most theoretical guarantees only provide an in-expectation analysis and only focus on optimization performance. In this paper, we provide high probability analysis in the non-convex setting and derive the optimization bound and the generalization bound simultaneously for popular stochastic optimization algorithms with gradient clipping, including stochastic gradient descent and its variants of momentum and adaptive stepsizes. With the gradient clipping, we study a heavy-tailed assumption that the gradients only have bounded $α$-th moments for some $α\in (1, 2]$, which is much weaker than the standard bounded second-moment assumption. Overall, our study provides a relatively complete picture for the theoretical guarantee of stochastic optimization algorithms with clipping.

Shaojie Li, Sheng Ouyang, Yong Liu

The theoretical analysis of spectral clustering mainly focuses on consistency, while there is relatively little research on its generalization performance. In this paper, we study the excess risk bounds of the popular spectral clustering algorithms: \emph{relaxed} RatioCut and \emph{relaxed} NCut. Firstly, we show that their excess risk bounds between the empirical continuous optimal solution and the population-level continuous optimal solution have a $\mathcal{O}(1/\sqrt{n})$ convergence rate, where $n$ is the sample size. Secondly, we show the fundamental quantity in influencing the excess risk between the empirical discrete optimal solution and the population-level discrete optimal solution. At the empirical level, algorithms can be designed to reduce this quantity. Based on our theoretical analysis, we propose two novel algorithms that can not only penalize this quantity, but also cluster the out-of-sample data without re-eigendecomposition on the overall sample. Experiments verify the effectiveness of the proposed algorithms.

Shaojie Li, Yunbei Xu

We address the fundamental question of why deep neural networks generalize by establishing a pointwise generalization theory for fully connected networks. This framework resolves long-standing barriers to characterizing the rich nonlinear feature-learning regime and builds a new statistical foundation for representation learning. For each trained model, we characterize the hypothesis via a pointwise Riemannian Dimension, derived from the eigenvalues of the learned feature representations across layers. This establishes a principled framework for deriving hypothesis-dependent, representation-aware generalization bounds. These bounds offer a systematic upgrade over approaches based on model size, products of norms, and infinite-width linearizations, yielding guarantees that are orders of magnitude tighter in both theory and experiment. Analytically, we identify the structural properties and mathematical principles that explain the tractability of deep networks. Empirically, the pointwise Riemannian Dimension exhibits substantial feature compression, decreases with increased over-parameterization, and captures the implicit bias of optimizers. Taken together, our results indicate that deep networks are mathematically tractable in practical regimes and that their generalization is sharply explained by pointwise, feature-spectrum-aware complexity.

Shaojie Li, Zhaoshuo Diao

The recent development of deep learning large models in medicine shows remarkable performance in medical image analysis and diagnosis, but their large number of parameters causes memory and inference latency challenges. Knowledge distillation offers a solution, but the slide-level gradients cannot be backpropagated for student model updates due to high-resolution pathological images and slide-level labels. This study presents an Efficient Fine-tuning on Compressed Models (EFCM) framework with two stages: unsupervised feature distillation and fine-tuning. In the distillation stage, Feature Projection Distillation (FPD) is proposed with a TransScan module for adaptive receptive field adjustment to enhance the knowledge absorption capability of the student model. In the slide-level fine-tuning stage, three strategies (Reuse CLAM, Retrain CLAM, and End2end Train CLAM (ETC)) are compared. Experiments are conducted on 11 downstream datasets related to three large medical models: RETFound for retina, MRM for chest X-ray, and BROW for histopathology. The experimental results demonstrate that the EFCM framework significantly improves accuracy and efficiency in handling slide-level pathological image problems, effectively addressing the challenges of deploying large medical models. Specifically, it achieves a 4.33% increase in ACC and a 5.2% increase in AUC compared to the large model BROW on the TCGA-NSCLC and TCGA-BRCA datasets. The analysis of model inference efficiency highlights the high efficiency of the distillation fine-tuning method.

Shaojie Li, Jie Wu, Xuefeng Xiao et al.

Recently, a series of algorithms have been explored for GAN compression, which aims to reduce tremendous computational overhead and memory usages when deploying GANs on resource-constrained edge devices. However, most of the existing GAN compression work only focuses on how to compress the generator, while fails to take the discriminator into account. In this work, we revisit the role of discriminator in GAN compression and design a novel generator-discriminator cooperative compression scheme for GAN compression, termed GCC. Within GCC, a selective activation discriminator automatically selects and activates convolutional channels according to a local capacity constraint and a global coordination constraint, which help maintain the Nash equilibrium with the lightweight generator during the adversarial training and avoid mode collapse. The original generator and discriminator are also optimized from scratch, to play as a teacher model to progressively refine the pruned generator and the selective activation discriminator. A novel online collaborative distillation scheme is designed to take full advantage of the intermediate feature of the teacher generator and discriminator to further boost the performance of the lightweight generator. Extensive experiments on various GAN-based generation tasks demonstrate the effectiveness and generalization of GCC. Among them, GCC contributes to reducing 80% computational costs while maintains comparable performance in image translation tasks. Our code and models are available at https://github.com/SJLeo/GCC.

Shaojie Li, Mingbao Lin, Yan Wang et al.

Existing online knowledge distillation approaches either adopt the student with the best performance or construct an ensemble model for better holistic performance. However, the former strategy ignores other students' information, while the latter increases the computational complexity during deployment. In this paper, we propose a novel method for online knowledge distillation, termed FFSD, which comprises two key components: Feature Fusion and Self-Distillation, towards solving the above problems in a unified framework. Different from previous works, where all students are treated equally, the proposed FFSD splits them into a leader student and a common student set. Then, the feature fusion module converts the concatenation of feature maps from all common students into a fused feature map. The fused representation is used to assist the learning of the leader student. To enable the leader student to absorb more diverse information, we design an enhancement strategy to increase the diversity among students. Besides, a self-distillation module is adopted to convert the feature map of deeper layers into a shallower one. Then, the shallower layers are encouraged to mimic the transformed feature maps of the deeper layers, which helps the students to generalize better. After training, we simply adopt the leader student, which achieves superior performance, over the common students, without increasing the storage or inference cost. Extensive experiments on CIFAR-100 and ImageNet demonstrate the superiority of our FFSD over existing works. The code is available at https://github.com/SJLeo/FFSD.

Mingbao Lin, Rongrong Ji, Shaojie Li et al.

Popular network pruning algorithms reduce redundant information by optimizing hand-crafted models, and may cause suboptimal performance and long time in selecting filters. We innovatively introduce adaptive exemplar filters to simplify the algorithm design, resulting in an automatic and efficient pruning approach called EPruner. Inspired by the face recognition community, we use a message passing algorithm Affinity Propagation on the weight matrices to obtain an adaptive number of exemplars, which then act as the preserved filters. EPruner breaks the dependency on the training data in determining the "important" filters and allows the CPU implementation in seconds, an order of magnitude faster than GPU based SOTAs. Moreover, we show that the weights of exemplars provide a better initialization for the fine-tuning. On VGGNet-16, EPruner achieves a 76.34%-FLOPs reduction by removing 88.80% parameters, with 0.06% accuracy improvement on CIFAR-10. In ResNet-152, EPruner achieves a 65.12%-FLOPs reduction by removing 64.18% parameters, with only 0.71% top-5 accuracy loss on ILSVRC-2012. Our code can be available at https://github.com/lmbxmu/EPruner.

Shaojie Li, Mingbao Lin, Yan Wang et al.

Generative Adversarial Networks (GANs) have been widely-used in image translation, but their high computation and storage costs impede the deployment on mobile devices. Prevalent methods for CNN compression cannot be directly applied to GANs due to the peculiarties of GAN tasks and the unstable adversarial training. To solve these, in this paper, we introduce a novel GAN compression method, termed DMAD, by proposing a Differentiable Mask and a co-Attention Distillation. The former searches for a light-weight generator architecture in a training-adaptive manner. To overcome channel inconsistency when pruning the residual connections, an adaptive cross-block group sparsity is further incorporated. The latter simultaneously distills informative attention maps from both the generator and discriminator of a pre-trained model to the searched generator, effectively stabilizing the adversarial training of our light-weight model. Experiments show that DMAD can reduce the Multiply Accumulate Operations (MACs) of CycleGAN by 13x and that of Pix2Pix by 4x while retaining a comparable performance against the full model. Our code can be available at https://github.com/SJLeo/DMAD.

Mingbao Lin, Liujuan Cao, Shaojie Li et al.

We propose a novel network pruning approach by information preserving of pre-trained network weights (filters). Network pruning with the information preserving is formulated as a matrix sketch problem, which is efficiently solved by the off-the-shelf Frequent Direction method. Our approach, referred to as FilterSketch, encodes the second-order information of pre-trained weights, which enables the representation capacity of pruned networks to be recovered with a simple fine-tuning procedure. FilterSketch requires neither training from scratch nor data-driven iterative optimization, leading to a several-orders-of-magnitude reduction of time cost in the optimization of pruning. Experiments on CIFAR-10 show that FilterSketch reduces 63.3% of FLOPs and prunes 59.9% of network parameters with negligible accuracy cost for ResNet-110. On ILSVRC-2012, it reduces 45.5% of FLOPs and removes 43.0% of parameters with only 0.69% accuracy drop for ResNet-50. Our code and pruned models can be found at https://github.com/lmbxmu/FilterSketch.

Shaojie Li, Xinqi Dong, Danqing Ma et al.

Mobile Internet user credit assessment is an important way for communication operators to establish decisions and formulate measures, and it is also a guarantee for operators to obtain expected benefits. However, credit evaluation methods have long been monopolized by financial industries such as banks and credit. As supporters and providers of platform network technology and network resources, communication operators are also builders and maintainers of communication networks. Internet data improves the user's credit evaluation strategy. This paper uses the massive data provided by communication operators to carry out research on the operator's user credit evaluation model based on the fusion LightGBM algorithm. First, for the massive data related to user evaluation provided by operators, key features are extracted by data preprocessing and feature engineering methods, and a multi-dimensional feature set with statistical significance is constructed; then, linear regression, decision tree, LightGBM, and other machine learning algorithms build multiple basic models to find the best basic model; finally, integrates Averaging, Voting, Blending, Stacking and other integrated algorithms to refine multiple fusion models, and finally establish the most suitable fusion model for operator user evaluation.

Danqing Ma, Shaojie Li, Bo Dang et al.

Transmission line detection technology is crucial for automatic monitoring and ensuring the safety of electrical facilities. The YOLOv5 series is currently one of the most advanced and widely used methods for object detection. However, it faces inherent challenges, such as high computational load on devices and insufficient detection accuracy. To address these concerns, this paper presents an enhanced lightweight YOLOv5 technique customized for mobile devices, specifically intended for identifying objects associated with transmission lines. The C3Ghost module is integrated into the convolutional network of YOLOv5 to reduce floating point operations per second (FLOPs) in the feature channel fusion process and improve feature expression performance. In addition, a FasterNet module is introduced to replace the c3 module in the YOLOv5 Backbone. The FasterNet module uses Partial Convolutions to process only a portion of the input channels, improving feature extraction efficiency and reducing computational overhead. To address the imbalance between simple and challenging samples in the dataset and the diversity of aspect ratios of bounding boxes, the wIoU v3 LOSS is adopted as the loss function. To validate the performance of the proposed approach, Experiments are conducted on a custom dataset of transmission line poles. The results show that the proposed model achieves a 1% increase in detection accuracy, a 13% reduction in FLOPs, and a 26% decrease in model parameters compared to the existing YOLOv5.In the ablation experiment, it was also discovered that while the Fastnet module and the CSghost module improved the precision of the original YOLOv5 baseline model, they caused a decrease in the mAP@.5-.95 metric. However, the improvement of the wIoUv3 loss function significantly mitigated the decline of the mAP@.5-.95 metric.

Shaojie Li, Haichen Qu, Xinqi Dong et al.

Exploring the application of deep learning technologies in the field of medical diagnostics, Magnetic Resonance Imaging (MRI) provides a unique perspective for observing and diagnosing complex neurodegenerative diseases such as Alzheimer Disease (AD). With advancements in deep learning, particularly in Convolutional Neural Networks (CNNs) and the Xception network architecture, we are now able to analyze and classify vast amounts of MRI data with unprecedented accuracy. The progress of this technology not only enhances our understanding of brain structural changes but also opens up new avenues for monitoring disease progression through non-invasive means and potentially allows for precise diagnosis in the early stages of the disease. This study aims to classify MRI images using deep learning models to identify different stages of Alzheimer Disease through a series of innovative data processing and model construction steps. Our experimental results show that the deep learning framework based on the Xception model achieved a 99.6% accuracy rate in the multi-class MRI image classification task, demonstrating its potential application value in assistive diagnosis. Future research will focus on expanding the dataset, improving model interpretability, and clinical validation to further promote the application of deep learning technology in the medical field, with the hope of bringing earlier diagnosis and more personalized treatment plans to Alzheimer Disease patients.

Bowei Zhu, Shaojie Li, Mingyang Yi et al.

Prior work (Klochkov $\&$ Zhivotovskiy, 2021) establishes at most $O\left(\log (n)/n\right)$ excess risk bounds via algorithmic stability for strongly-convex learners with high probability. We show that under the similar common assumptions -- - Polyak-Lojasiewicz condition, smoothness, and Lipschitz continous for losses -- - rates of $O\left(\log^2(n)/n^2\right)$ are at most achievable. To our knowledge, our analysis also provides the tightest high-probability bounds for gradient-based generalization gaps in nonconvex settings.

Bowei Zhu, Shaojie Li, Yong Liu

Minimax problems have achieved success in machine learning such as adversarial training, robust optimization, reinforcement learning. For theoretical analysis, current optimal excess risk bounds, which are composed by generalization error and optimization error, present 1/n-rates in strongly-convex-strongly-concave (SC-SC) settings. Existing studies mainly focus on minimax problems with specific algorithms for optimization error, with only a few studies on generalization performance, which limit better excess risk bounds. In this paper, we study the generalization bounds measured by the gradients of primal functions using uniform localized convergence. We obtain a sharper high probability generalization error bound for nonconvex-strongly-concave (NC-SC) stochastic minimax problems. Furthermore, we provide dimension-independent results under Polyak-Lojasiewicz condition for the outer layer. Based on our generalization error bound, we analyze some popular algorithms such as empirical saddle point (ESP), gradient descent ascent (GDA) and stochastic gradient descent ascent (SGDA). We derive better excess primal risk bounds with further reasonable assumptions, which, to the best of our knowledge, are n times faster than exist results in minimax problems.

Shaojie Li, Yong Liu

Pairwise learning is receiving increasing attention since it covers many important machine learning tasks, e.g., metric learning, AUC maximization, and ranking. Investigating the generalization behavior of pairwise learning is thus of significance. However, existing generalization analysis mainly focuses on the convex objective functions, leaving the nonconvex learning far less explored. Moreover, the current learning rates derived for generalization performance of pairwise learning are mostly of slower order. Motivated by these problems, we study the generalization performance of nonconvex pairwise learning and provide improved learning rates. Specifically, we develop different uniform convergence of gradients for pairwise learning under different assumptions, based on which we analyze empirical risk minimizer, gradient descent, and stochastic gradient descent pairwise learning. We first successfully establish learning rates for these algorithms in a general nonconvex setting, where the analysis sheds insights on the trade-off between optimization and generalization and the role of early-stopping. We then investigate the generalization performance of nonconvex learning with a gradient dominance curvature condition. In this setting, we derive faster learning rates of order $\mathcal{O}(1/n)$, where $n$ is the sample size. Provided that the optimal population risk is small, we further improve the learning rates to $\mathcal{O}(1/n^2)$, which, to the best of our knowledge, are the first $\mathcal{O}(1/n^2)$-type of rates for pairwise learning, no matter of convex or nonconvex learning. Overall, we systematically analyzed the generalization performance of nonconvex pairwise learning.

Shaojie Li, Yong Liu

Generalization performance of stochastic optimization stands a central place in learning theory. In this paper, we investigate the excess risk performance and towards improved learning rates for two popular approaches of stochastic optimization: empirical risk minimization (ERM) and stochastic gradient descent (SGD). Although there exists plentiful generalization analysis of ERM and SGD for supervised learning, current theoretical understandings of ERM and SGD either have stronger assumptions in convex learning, e.g., strong convexity, or show slow rates and less studied in nonconvex learning. Motivated by these problems, we aim to provide improved rates under milder assumptions in convex learning and derive faster rates in nonconvex learning. It is notable that our analysis span two popular theoretical viewpoints: \emph{stability} and \emph{uniform convergence}. Specifically, in stability regime, we present high probability learning rates of order $\mathcal{O} (1/n)$ w.r.t. the sample size $n$ for ERM and SGD with milder assumptions in convex learning and similar high probability rates of order $\mathcal{O} (1/n)$ in nonconvex learning, rather than in expectation. Furthermore, this type of learning rate is improved to faster order $\mathcal{O} (1/n^2)$ in uniform convergence regime. To our best knowledge, for ERM and SGD, the learning rates presented in this paper are all state-of-the-art.