Unlike the conventional Knowledge Distillation (KD), Self-KD allows a network to learn knowledge from itself without any guidance from extra networks. This paper proposes to perform Self-KD from image Mixture (MixSKD), which integrates these two techniques into a unified framework. MixSKD mutually distills feature maps and probability distributions between the random pair of original images and their mixup images in a meaningful way. Therefore, it guides the network to learn cross-image knowledge by modelling supervisory signals from mixup images. Moreover, we construct a self-teacher network by aggregating multi-stage feature maps for providing soft labels to supervise the backbone classifier, further improving the efficacy of self-boosting. Experiments on image classification and transfer learning to object detection and semantic segmentation demonstrate that MixSKD outperforms other state-of-the-art Self-KD and data augmentation methods. The code is available at https://github.com/winycg/Self-KD-Lib.
To boost the performance, deep neural networks require deeper or wider network structures that involve massive computational and memory costs. To alleviate this issue, the self-knowledge distillation method regularizes the model by distilling the internal knowledge of the model itself. Conventional self-knowledge distillation methods require additional trainable parameters or are dependent on the data. In this paper, we propose a simple and effective self-knowledge distillation method using a dropout (SD-Dropout). SD-Dropout distills the posterior distributions of multiple models through a dropout sampling. Our method does not require any additional trainable modules, does not rely on data, and requires only simple operations. Furthermore, this simple method can be easily combined with various self-knowledge distillation approaches. We provide a theoretical and experimental analysis of the effect of forward and reverse KL-divergences in our work. Extensive experiments on various vision tasks, i.e., image classification, object detection, and distribution shift, demonstrate that the proposed method can effectively improve the generalization of a single network. Further experiments show that the proposed method also improves calibration performance, adversarial robustness, and out-of-distribution detection ability.
The global outbreak of the Coronavirus 2019 (COVID-19) has overloaded worldwide healthcare systems. Computer-aided diagnosis for COVID-19 fast detection and patient triage is becoming critical. This paper proposes a novel self-knowledge distillation based self-supervised learning method for COVID-19 detection from chest X-ray images. Our method can use self-knowledge of images based on similarities of their visual features for self-supervised learning. Experimental results show that our method achieved an HM score of 0.988, an AUC of 0.999, and an accuracy of 0.957 on the largest open COVID-19 chest X-ray dataset.
Knowledge Graph Embedding (KGE), which projects entities and relations into continuous vector spaces, has garnered significant attention. Although high-dimensional KGE methods offer better performance, they come at the expense of significant computation and memory overheads. Decreasing embedding dimensions significantly deteriorates model performance. While several recent efforts utilize knowledge distillation or non-Euclidean representation learning to augment the effectiveness of low-dimensional KGE, they either necessitate a pre-trained high-dimensional teacher model or involve complex non-Euclidean operations, thereby incurring considerable additional computational costs. To address this, this work proposes Confidence-aware Self-Knowledge Distillation (CSD) that learns from the model itself to enhance KGE in a low-dimensional space. Specifically, CSD extracts knowledge from embeddings in previous iterations, which would be utilized to supervise the learning of the model in the next iterations. Moreover, a specific semantic module is developed to filter reliable knowledge by estimating the confidence of previously learned embeddings. This straightforward strategy bypasses the need for time-consuming pre-training of teacher models and can be integrated into various KGE methods to improve their performance. Our comprehensive experiments on six KGE backbones and four datasets underscore the effectiveness of the proposed CSD.
Weakly supervised object detection (WSOD), which is an effective way to train an object detection model using only image-level annotations, has attracted considerable attention from researchers. However, most of the existing methods, which are based on multiple instance learning (MIL), tend to localize instances to the discriminative parts of salient objects instead of the entire content of all objects. In this paper, we propose a WSOD framework called the Spatial Likelihood Voting with Self-knowledge Distillation Network (SLV-SD Net). In this framework, we introduce a spatial likelihood voting (SLV) module to converge region proposal localization without bounding box annotations. Specifically, in every iteration during training, all the region proposals in a given image act as voters voting for the likelihood of each category in the spatial dimensions. After dilating the alignment on the area with large likelihood values, the voting results are regularized as bounding boxes, which are then used for the final classification and localization. Based on SLV, we further propose a self-knowledge distillation (SD) module to refine the feature representations of the given image. The likelihood maps generated by the SLV module are used to supervise the feature learning of the backbone network, encouraging the network to attend to wider and more diverse areas of the image. Extensive experiments on the PASCAL VOC 2007/2012 and MS-COCO datasets demonstrate the excellent performance of SLV-SD Net. In addition, SLV-SD Net produces new state-of-the-art results on these benchmarks.
Neural Machine Translation(NMT) models are usually trained via unidirectional decoder which corresponds to optimizing one-step-ahead prediction. However, this kind of unidirectional decoding framework may incline to focus on local structure rather than global coherence. To alleviate this problem, we propose a novel method, Self-Knowledge Distillation with Bidirectional Decoder for Neural Machine Translation(SBD-NMT). We deploy a backward decoder which can act as an effective regularization method to the forward decoder. By leveraging the backward decoder's information about the longer-term future, distilling knowledge learned in the backward decoder can encourage auto-regressive NMT models to plan ahead. Experiments show that our method is significantly better than the strong Transformer baselines on multiple machine translation data sets.
Pothole classification has become an important task for road inspection vehicles to save drivers from potential car accidents and repair bills. Given the limited computational power and fixed number of training epochs, we propose iterative self knowledge distillation (ISKD) to train lightweight pothole classifiers. Designed to improve both the teacher and student models over time in knowledge distillation, ISKD outperforms the state-of-the-art self knowledge distillation method on three pothole classification datasets across four lightweight network architectures, which supports that self knowledge distillation should be done iteratively instead of just once. The accuracy relation between the teacher and student models shows that the student model can still benefit from a moderately trained teacher model. Implying that better teacher models generally produce better student models, our results justify the design of ISKD. In addition to pothole classification, we also demonstrate the efficacy of ISKD on six additional datasets associated with generic classification, fine-grained classification, and medical imaging application, which supports that ISKD can serve as a general-purpose performance booster without the need of a given teacher model and extra trainable parameters.
Knowledge distillation aims at transferring the knowledge from a large teacher model to a small student model with great improvements of the performance of the student model. Therefore, the student network can replace the teacher network to deploy on low-resource devices since the higher performance, lower number of parameters and shorter inference time. Self-knowledge distillation (SKD) attracts a great attention recently that a student model itself is a teacher model distilling knowledge from. To the best of our knowledge, self knowledge distillation can be divided into two main streams: data augmentation and refined knowledge auxiliary. In this paper, we purpose a novel SKD method in a different way from the main stream methods. Our method distills knowledge from multilevel abstraction features. Experiments and ablation studies show its great effectiveness and generalization on various kinds of tasks with various kinds of model structures. Our codes have been released on GitHub.
Sequential recommendation systems are integral to discerning temporal user preferences. Yet, the task of learning from abbreviated user interaction sequences poses a notable challenge. Data augmentation has been identified as a potent strategy to enhance the informational richness of these sequences. Traditional augmentation techniques, such as item randomization, may disrupt the inherent temporal dynamics. Although recent advancements in reverse chronological pseudo-item generation have shown promise, they can introduce temporal discrepancies when assessed in a natural chronological context. In response, we introduce a sophisticated approach, Bidirectional temporal data Augmentation with pre-training (BARec). Our approach leverages bidirectional temporal augmentation and knowledge-enhanced fine-tuning to synthesize authentic pseudo-prior items that retain user preferences and capture deeper item semantic correlations, thus boosting the model's expressive power. Our comprehensive experimental analysis on five benchmark datasets confirms the superiority of BARec across both short and elongated sequence contexts. Moreover, theoretical examination and case study offer further insight into the model's logical processes and interpretability. The source code for our study is publicly available at https://github.com/juyongjiang/BARec.
Object detection has achieved promising performance on clean datasets, but how to achieve better tradeoff between the adversarial robustness and clean precision is still under-explored. Adversarial training is the mainstream method to improve robustness, but most of the works will sacrifice clean precision to gain robustness than standard training. In this paper, we propose Unified Decoupled Feature Alignment (UDFA), a novel fine-tuning paradigm which achieves better performance than existing methods, by fully exploring the combination between self-knowledge distillation and adversarial training for object detection. We first use decoupled fore/back-ground features to construct self-knowledge distillation branch between clean feature representation from pretrained detector (served as teacher) and adversarial feature representation from student detector. Then we explore the self-knowledge distillation from a new angle by decoupling original branch into a self-supervised learning branch and a new self-knowledge distillation branch. With extensive experiments on the PASCAL-VOC and MS-COCO benchmarks, the evaluation results show that UDFA can surpass the standard training and state-of-the-art adversarial training methods for object detection. For example, compared with teacher detector, our approach on GFLV2 with ResNet-50 improves clean precision by 2.2 AP on PASCAL-VOC; compared with SOTA adversarial training methods, our approach improves clean precision by 1.6 AP, while improving adversarial robustness by 0.5 AP. Our code will be available at https://github.com/grispeut/udfa.
Knowledge distillation usually transfers the knowledge from a pre-trained cumbersome teacher network to a compact student network, which follows the classical teacher-teaching-student paradigm. Based on this paradigm, previous methods mostly focus on how to efficiently train a better student network for deployment. Different from the existing practices, in this paper, we propose a novel student-helping-teacher formula, Teacher Evolution via Self-Knowledge Distillation (TESKD), where the target teacher (for deployment) is learned with the help of multiple hierarchical students by sharing the structural backbone. The diverse feedback from multiple students allows the teacher to improve itself through the shared feature representations. The effectiveness of our proposed framework is demonstrated by extensive experiments with various network settings on two standard benchmarks including CIFAR-100 and ImageNet. Notably, when trained together with our proposed method, ResNet-18 achieves 79.15% and 71.14% accuracy on CIFAR-100 and ImageNet, outperforming the baseline results by 4.74% and 1.43%, respectively. The code is available at: https://github.com/zhengli427/TESKD.
In federated learning, a strong global model is collaboratively learned by aggregating clients' locally trained models. Although this precludes the need to access clients' data directly, the global model's convergence often suffers from data heterogeneity. This study starts from an analogy to continual learning and suggests that forgetting could be the bottleneck of federated learning. We observe that the global model forgets the knowledge from previous rounds, and the local training induces forgetting the knowledge outside of the local distribution. Based on our findings, we hypothesize that tackling down forgetting will relieve the data heterogeneity problem. To this end, we propose a novel and effective algorithm, Federated Not-True Distillation (FedNTD), which preserves the global perspective on locally available data only for the not-true classes. In the experiments, FedNTD shows state-of-the-art performance on various setups without compromising data privacy or incurring additional communication costs.
Previous studies have verified that the functionality of black-box models can be stolen with full probability outputs. However, under the more practical hard-label setting, we observe that existing methods suffer from catastrophic performance degradation. We argue this is due to the lack of rich information in the probability prediction and the overfitting caused by hard labels. To this end, we propose a novel hard-label model stealing method termed \emph{black-box dissector}, which consists of two erasing-based modules. One is a CAM-driven erasing strategy that is designed to increase the information capacity hidden in hard labels from the victim model. The other is a random-erasing-based self-knowledge distillation module that utilizes soft labels from the substitute model to mitigate overfitting. Extensive experiments on four widely-used datasets consistently demonstrate that our method outperforms state-of-the-art methods, with an improvement of at most $8.27\%$. We also validate the effectiveness and practical potential of our method on real-world APIs and defense methods. Furthermore, our method promotes other downstream tasks, \emph{i.e.}, transfer adversarial attacks.
End-to-end automatic speech recognition (ASR), unlike conventional ASR, does not have modules to learn the semantic representation from speech encoder. Moreover, the higher frame-rate of speech representation prevents the model to learn the semantic representation properly. Therefore, the models that are constructed by the lower frame-rate of speech encoder lead to better performance. For Transformer-based ASR, the lower frame-rate is not only important for learning better semantic representation but also for reducing the computational complexity due to the self-attention mechanism which has O(n^2) order of complexity in both training and inference. In this paper, we propose a Transformer-based ASR model with the time reduction layer, in which we incorporate time reduction layer inside transformer encoder layers in addition to traditional sub-sampling methods to input features that further reduce the frame-rate. This can help in reducing the computational cost of the self-attention process for training and inference with performance improvement. Moreover, we introduce a fine-tuning approach for pre-trained ASR models using self-knowledge distillation (S-KD) which further improves the performance of our ASR model. Experiments on LibriSpeech datasets show that our proposed methods outperform all other Transformer-based ASR systems. Furthermore, with language model (LM) fusion, we achieve new state-of-the-art word error rate (WER) results for Transformer-based ASR models with just 30 million parameters trained without any external data.
Knowledge distillation is a method of transferring the knowledge from a pretrained complex teacher model to a student model, so a smaller network can replace a large teacher network at the deployment stage. To reduce the necessity of training a large teacher model, the recent literatures introduced a self-knowledge distillation, which trains a student network progressively to distill its own knowledge without a pretrained teacher network. While Self-knowledge distillation is largely divided into a data augmentation based approach and an auxiliary network based approach, the data augmentation approach looses its local information in the augmentation process, which hinders its applicability to diverse vision tasks, such as semantic segmentation. Moreover, these knowledge distillation approaches do not receive the refined feature maps, which are prevalent in the object detection and semantic segmentation community. This paper proposes a novel self-knowledge distillation method, Feature Refinement via Self-Knowledge Distillation (FRSKD), which utilizes an auxiliary self-teacher network to transfer a refined knowledge for the classifier network. Our proposed method, FRSKD, can utilize both soft label and feature-map distillations for the self-knowledge distillation. Therefore, FRSKD can be applied to classification, and semantic segmentation, which emphasize preserving the local information. We demonstrate the effectiveness of FRSKD by enumerating its performance improvements in diverse tasks and benchmark datasets. The implemented code is available at https://github.com/MingiJi/FRSKD.
Knowledge distillation is classically a procedure where a neural network is trained on the output of another network along with the original targets in order to transfer knowledge between the architectures. The special case of self-distillation, where the network architectures are identical, has been observed to improve generalization accuracy. In this paper, we consider an iterative variant of self-distillation in a kernel regression setting, in which successive steps incorporate both model outputs and the ground-truth targets. This allows us to provide the first theoretical results on the importance of using the weighted ground-truth targets in self-distillation. Our focus is on fitting nonlinear functions to training data with a weighted mean square error objective function suitable for distillation, subject to $\ell_2$ regularization of the model parameters. We show that any such function obtained with self-distillation can be calculated directly as a function of the initial fit, and that infinite distillation steps yields the same optimization problem as the original with amplified regularization. Furthermore, we provide a closed form solution for the optimal choice of weighting parameter at each step, and show how to efficiently estimate this weighting parameter for deep learning and significantly reduce the computational requirements compared to a grid search.
The quantization of deep neural networks (QDNNs) has been actively studied for deployment in edge devices. Recent studies employ the knowledge distillation (KD) method to improve the performance of quantized networks. In this study, we propose stochastic precision ensemble training for QDNNs (SPEQ). SPEQ is a knowledge distillation training scheme; however, the teacher is formed by sharing the model parameters of the student network. We obtain the soft labels of the teacher by changing the bit precision of the activation stochastically at each layer of the forward-pass computation. The student model is trained with these soft labels to reduce the activation quantization noise. The cosine similarity loss is employed, instead of the KL-divergence, for KD training. As the teacher model changes continuously by random bit-precision assignment, it exploits the effect of stochastic ensemble KD. SPEQ outperforms the existing quantization training methods in various tasks, such as image classification, question-answering, and transfer learning without the need for cumbersome teacher networks.
In this paper we apply self-knowledge distillation to text summarization which we argue can alleviate problems with maximum-likelihood training on single reference and noisy datasets. Instead of relying on one-hot annotation labels, our student summarization model is trained with guidance from a teacher which generates smoothed labels to help regularize training. Furthermore, to better model uncertainty during training, we introduce multiple noise signals for both teacher and student models. We demonstrate experimentally on three benchmarks that our framework boosts the performance of both pretrained and non-pretrained summarizers achieving state-of-the-art results.
Inspired by the strong correlation between the Label Smoothing Regularization(LSR) and Knowledge distillation(KD), we propose an algorithm LsrKD for training boost by extending the LSR method to the KD regime and applying a softer temperature. Then we improve the LsrKD by a Teacher Correction(TC) method, which manually sets a constant larger proportion for the right class in the uniform distribution teacher. To further improve the performance of LsrKD, we develop a self-distillation method named Memory-replay Knowledge Distillation (MrKD) that provides a knowledgeable teacher to replace the uniform distribution one in LsrKD. The MrKD method penalizes the KD loss between the current model's output distributions and its copies' on the training trajectory. By preventing the model learning so far from its historical output distribution space, MrKD can stabilize the learning and find a more robust minimum. Our experiments show that LsrKD can improve LSR performance consistently at no cost, especially on several deep neural networks where LSR is ineffectual. Also, MrKD can significantly improve single model training. The experiment results confirm that the TC can help LsrKD and MrKD to boost training, especially on the networks they are failed. Overall, LsrKD, MrKD, and their TC variants are comparable to or outperform the LSR method, suggesting the broad applicability of these KD methods.
The generalization capability of deep neural networks has been substantially improved by applying a wide spectrum of regularization methods, e.g., restricting function space, injecting randomness during training, augmenting data, etc. In this work, we propose a simple yet effective regularization method named progressive self-knowledge distillation (PS-KD), which progressively distills a model's own knowledge to soften hard targets (i.e., one-hot vectors) during training. Hence, it can be interpreted within a framework of knowledge distillation as a student becomes a teacher itself. Specifically, targets are adjusted adaptively by combining the ground-truth and past predictions from the model itself. We show that PS-KD provides an effect of hard example mining by rescaling gradients according to difficulty in classifying examples. The proposed method is applicable to any supervised learning tasks with hard targets and can be easily combined with existing regularization methods to further enhance the generalization performance. Furthermore, it is confirmed that PS-KD achieves not only better accuracy, but also provides high quality of confidence estimates in terms of calibration as well as ordinal ranking. Extensive experimental results on three different tasks, image classification, object detection, and machine translation, demonstrate that our method consistently improves the performance of the state-of-the-art baselines. The code is available at https://github.com/lgcnsai/PS-KD-Pytorch.