DDDG: A dual bi-directional knowledge distillation method with generative self-supervised pre-training and its hardware implementation on SoC for ECG

Current Knowledge Distillation (KD) methods for semantic segmentation often guide the student to mimic the teacher's structured information generated from individual data samples. However, they ignore the global semantic relations among pixels across various images that are valuable for KD. This paper proposes a novel Cross-Image Relational KD (CIRKD), which focuses on transferring structured pixel-to-pixel and pixel-to-region relations among the whole images. The motivation is that a good teacher network could construct a well-structured feature space in terms of global pixel dependencies. CIRKD makes the student mimic better structured semantic relations from the teacher, thus improving the segmentation performance. Experimental results over Cityscapes, CamVid and Pascal VOC datasets demonstrate the effectiveness of our proposed approach against state-of-the-art distillation methods. The code is available at https://github.com/winycg/CIRKD.

Maximizing discrimination capability of knowledge distillation with energy function

A single model usually cannot learn all the appropriate features with limited data, thus leading to poor performance when test data are used. To improve model performance, we propose a teacher-student collaborative knowledge distillation (TSKD) method based on knowledge distillation and self-distillation. The method consists of two parts: learning in the teacher network and self-teaching in the student network. Learning in the teacher network allows the student network to use knowledge from the teacher network. Self-teaching in the student network is to build a multi-exit network based on self-distillation and provide deep features as supervised information for training. In the inference stage, we use ensembles to vote on the classification results of multiple sub-models in the student network. The experimental results demonstrate the superior performance of our method compared with a traditional knowledge distillation method and a self-distillation-based multi-exit network.

Knowledge distillation is a deep learning method that mimics the way that humans teach, i.e., a teacher network is used to guide the training of a student one. Knowledge distillation can generate an efficient student network to facilitate deployment in resource-constrained edge computing devices. Existing studies have typically mined knowledge from a teacher network and transferred it to a student one. The latter can only passively receive knowledge but cannot understand how the former acquires the knowledge, thus limiting the latter’s performance improvement. Inspired by the old Chinese saying “Give a man a fish and you feed him for a day; teach a man how to fish and you feed him for a lifetime,” this work proposes a Skill-transferring Knowledge Distillation (SKD) method to boost a student network’s ability to create new valuable knowledge. SKD consists of two main meta-learning networks: Teacher Behavior Teaching and Teacher Experience Teaching. The former captures the process of a teacher network’s learning behavior in the hidden layers and can predict the teacher network’s subsequent behavior based on previous ones. The latter models the optimal empirical knowledge of a teacher network’s output layer at each learning stage. With their help, a teacher network can provide its actions to a student one in the subsequent behavior and its optimal empirical knowledge in the current stage. The overall performance of SKD is verified through its application to multiple object recognition tasks and comparison with the state of the art.

This study addresses the problem of the unsupervised pre‐training of video representation learning. The authors' focus is on two common approaches: knowledge distillation and self‐supervised learning. The insight provided is that knowledge distillation and the rapidly advancing self‐supervised technique can be mutually beneficial. Combining these two approaches, a unified framework of self‐supervised learning and image‐based distillation (SSID) for unsupervised video pre‐training is proposed. The effectiveness of SSID in comparison to both image‐based distillation methods and the existing self‐supervised pre‐training baseline is demonstrated. In particular, the authors' model leverages three signals from the unlabelled data. First, the authors distil from the classifier of a 2D pre‐trained model as a soft label. To regularize the training process, the authors then build a novel positive pair of contrastive learning on the representation of the 2D/3D model. Finally, a self‐supervised pretext task is introduced to enhance the authors' model to become aware of the temporal evolution. The authors' experiment results showed that the learnt features achieved the best performance when transferred to action recognition tasks on UCF101 and HMDB51, reaching increases of 2.4% and 1.9% compared to the existing unsupervised pre‐training model, respectively.

Knowledge Distillation (KD) aims to distill the dark knowledge of a high-powered teacher network into a student network, which can improve the capacity of student network and has been successfully applied to semantic segmentation. However, the standard knowledge distillation approaches merely represent the supervisory signal of teacher network as the dark knowledge, while ignoring the impact of network architecture during distillation. In this paper, we found that the student network with a more similar architecture against the teacher network obtains more performance gain from distillation. Therefore, a more generalized paradigm for knowledge distillation is to distill both the soft-label and the structure of the teacher network. We propose a novel Structural Distillation (SD) method which introduces the structural similarity constraints into vanilla knowledge distillation. We leverage Neural Architecture Search technique to search optimal student structure for semantic segmentation from a well-designed search space, which mimics the given teacher both in terms of soft-label and network structure. Experiment results make clear that our proposed method outperforms both the NAS with conventional Knowledge Distillation and human-designed methods, and achieves sota performance on the Cityscapes dataset under various platform-aware latency constraints. Furthermore, the best architecture discovered on Cityscapes also transfers well to the PASCAL VOC2012 dataset.

Graph Neural Networks (GNNs) have shown remarkable effectiveness across various applications, but their computational complexity poses significant scalability challenges. To this end, GNN-to-MLP Knowledge Distillation (KD) methods transfer relational inductive biases from GNNs to MLPs, equipping MLPs with graph-aware capabilities that rival or even surpass those of their teacher GNNs. However, a theoretical foundation for understanding GNN-to-MLP KD is still missing. In this paper, we provide a theoretical analysis of how knowledge distillation unlocks the potential of MLPs for graph tasks from the perspective of training dynamics. We demonstrate that label alignment in KD fundamentally reshapes the Neural Tangent Kernel (NTK) matrix of student MLPs, enabling them to learn the teacher model’s implicit graph bias. We further investigate finer-grained distillation paradigms and reveal that conventional layer-wise output alignment fails to effectively align the deep-layer graph propagation outcomes. To address this, we propose Dual-Stream Aligned MLP (DA-MLP), which incorporates complementary graph filters in a dual-stream architecture. This approach simultaneously enhances feature space dimensionality for improved representation alignment and preserves graph signals across different frequency bands. Comprehensive experiments on seven benchmark datasets validate that DA-MLP can be seamlessly integrated into existing knowledge distillation frameworks for performance enhancements in both transductive and inductive settings.

Medical image classification: Knowledge transfer via residual U-Net and vision transformer-based teacher-student model with knowledge distillation

VFT: A versatile fine-tuning scheme based on feature distribution-aware knowledge distillation for lightweight convolutional neural networks

We present Knowledge Distillation with Meta Learning (MetaDistil), a simple yet effective alternative to traditional knowledge distillation (KD) methods where the teacher model is fixed during training. We show the teacher network can learn to better transfer knowledge to the student network (i.e., learning to teach) with the feedback from the performance of the distilled student network in a meta learning framework. Moreover, we introduce a pilot update mechanism to improve the alignment between the inner-learner and meta-learner in meta learning algorithms that focus on an improved inner-learner. Experiments on various benchmarks show that MetaDistil can yield significant improvements compared with traditional KD algorithms and is less sensitive to the choice of different student capacity and hyperparameters, facilitating the use of KD on different tasks and models.

Improving adversarial robustness using knowledge distillation guided by attention information bottleneck

Knowledge distillation has been used to transfer knowledge learned by a sophisticated model (teacher) to a simpler model (student). This technique is widely used to compress model complexity. However, in most applications the compressed student model suffers from an accuracy gap with its teacher. We propose extracurricular learning, a novel knowledge distillation method, that bridges this gap by (1) modeling student and teacher output distributions; (2) sampling examples from an approximation to the underlying data distribution; and (3) matching student and teacher output distributions over this extended set including uncertain samples. We conduct rigorous evaluations on regression and classification tasks and show that compared to the standard knowledge distillation, extracurricular learning reduces the gap by 46% to 68%. This leads to major accuracy improvements compared to the empirical risk minimization-based training for various recent neural network architectures: 16% regression error reduction on the MPIIGaze dataset, +3.4% to +9.1% improvement in top-1 classification accuracy on the CIFAR100 dataset, and +2.9% top-1 improvement on the ImageNet dataset.

Knowledge distillation has emerged as a highly effective method for bridging the representation discrepancy between large-scale models and lightweight models. Prevalent approaches involve leveraging appropriate metrics to minimize the divergence or distance between the knowledge extracted from the teacher model and the knowledge learned by the student model. Centered Kernel Alignment (CKA) is widely used to measure representation similarity and has been applied in several knowledge distillation methods. However, these methods are complex and fail to uncover the essence of CKA, thus not answering the question of how to use CKA to achieve simple and effective distillation properly. This paper first provides a theoretical perspective to illustrate the effectiveness of CKA, which decouples CKA to the upper bound of Maximum Mean Discrepancy (MMD) and a constant term. Drawing from this, we propose a novel Relation-Centered Kernel Alignment (RCKA) framework, which practically establishes a connection between CKA and MMD. Furthermore, we dynamically customize the application of CKA based on the characteristics of each task, with less computational source yet comparable performance than the previous methods. The extensive experiments on the CIFAR-100, ImageNet-1k, and MS-COCO demonstrate that our method achieves state-of-the-art performance on almost all teacher-student pairs for image classification and object detection, validating the effectiveness of our approaches. Our code is available in https://github.com/Klayand/PCKA.

While deep models have shown promising performance in medical image segmentation, they heavily rely on a large amount of well-annotated data, which is difficult to access, especially in clinical practice. On the other hand, high-accuracy deep models usually come in large model sizes, limiting their employment in real scenarios. In this work, we propose a novel asymmetric co-teacher framework, ACT-Net, to alleviate the burden on both expensive annotations and computational costs for semi-supervised knowledge distillation. We advance teacher-student learning with a co-teacher network to facilitate asymmetric knowledge distillation from large models to small ones by alternating student and teacher roles, obtaining tiny but accurate models for clinical employment. To verify the effectiveness of our ACT-Net, we employ the ACDC dataset for cardiac substructure segmentation in our experiments. Extensive experimental results demonstrate that ACT-Net outperforms other knowledge distillation methods and achieves lossless segmentation performance with 250x fewer parameters.

<title>Abstract</title> The advent of wearable sensor technologies has transformed sports science and personal health monitoring, but real-time Human Activity Recognition (HAR) remains a challenge, especially in dynamic environments. Existing approaches often struggle to adapt to rapid activity transitions, handle multimodal sensor noise, and optimize for computational efficiency on resource-constrained wearable devices. In this work, we introduce an innovative system that overcomes these limitations through a novel integration of deep learning models and sensor fusion techniques. At its core, our framework features the Attention-Gated Spatiotemporal Fusion Network (AG-SFN), a deep neural network that combines Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks with a unique attention mechanism to dynamically prioritize the most relevant features from multimodal time-series data for accurate activity classification. To address sensor noise and improve data integrity, we propose the Adaptive Multimodal Kalman Fusion (AMKF) algorithm, which adapts the filtering process based on the motion intensity and sensor variance in real time. To ensure the efficient deployment of this model on wearable devices, we introduce a Dynamic Channel Pruning with Knowledge Distillation (DCP-KD) method, which significantly reduces the model size and computational load by pruning less relevant features and distilling knowledge from a large teacher model to a smaller student model. Our system was rigorously evaluated on a custom dataset of 40 participants performing six distinct sport-related activities, achieving a state-of-the-art activity recognition accuracy of 98.6%, outperforming conventional models by a significant margin. The system also demonstrated exceptional real-time performance with heart rate monitoring error reduced to 0.85 beats per minute (bpm), latency under 90 ms, and a 35% reduction in energy consumption. This research establishes a new benchmark for the development of efficient and adaptive wearable systems for real-time sports health monitoring, offering a robust solution for next-generation wearable technologies with practical applications in sports, fitness, and healthcare.