Knowledge Distillation with Adversarial Samples Supporting Decision Boundary

Recent studies have shown that robustness to adversarial attacks can be transferred across deep neural networks. In other words, we can make a weak model more robust with the help of a strong teacher model. In this paper, we ask if models can &#x201C;learn together&#x201D; and &#x201C;teach each other&#x201D; to achieve better robustness instead of learning from a static teacher.We study how interactions among models affect robustness via knowledge distillation. We propose mutual adversarial training (MAT), in which multiple models are trained together and share the knowledge of adversarial examples to achieve improved robustness. MAT allows robust models to explore a larger space of adversarial samples and find more robust feature spaces and decision boundaries. Through extensive experiments on the CIFAR-10, CIFAR-100, and mini-ImageNet datasets, we demonstrate that MAT can effectively improve model robustness and outperform state-of-the-art methods under white-box attacks. In addition, we show that MAT can also mitigate the robustness trade-off among different perturbation types. Specially, we train specialist models that learn to defend a specific perturbation type and a generalist model that learns to defend multiple perturbation types by learning from the specialists, which brings as much as 13.4% accuracy gain to AT baselines against the union of <i>l</i>&#x221E;, <i>l</i>2, and <i>l</i>1 attacks. Our results show the superiority of the proposed method and demonstrate that collaborative learning is an effective strategy for designing robust models.

Knowledge distillation based on decision boundary instances generated by DBI-GAN

Protein-protein interaction (PPI) plays an extremely important role in almost all life activities. The study of PPI has always been an important issue in the field of biomedicine. Extracting PPI information from the literature can provide important references for related research. In order to build an automated PPI extraction system, labeled corpora are required. However, annotated corpora are very limited, and annotating corpora is a time-consuming, labor-intensive, and costly task. The amount of unlabeled data is huge and easy to obtain, so it is of great significance to apply semi-supervised learning to PPI extraction. Existing semi-supervised methods cannot make good use of the information in unlabeled data. Therefore, a semi-supervised PPI extraction method based on knowledge distillation and virtual adversarial training is proposed in this work, which enhances the modeling ability of the target model for text features through knowledge distillation, and conducts semi-supervised learning through virtual adversarial training to better utilize a large number of unlabeled data. Experiments on public datasets for PPI extraction show that the proposed method can achieve competitive performance while using only a small amount of labeled data.

Knowledge distillation, in which the parameter values learned in a large teacher network are transferred to a smaller student network, is a popular and effective network compression method. Recently, researchers have proposed methods to improve the performance of a student network by using a Generative Adverserial Network (GAN). However, because a GAN is an architecture that is ideally used to create realistic synthetic images, a pure GAN architecture may not be ideally suited for knowledge distillation. In knowledge distillation for image signal processing, synthetic images do not need to be realistic, but instead should include features that help the training of the student network. In the proposed Generative Image Processing (GIP) method, this is accomplished by using only the generator portion of a GAN and utilizing special techniques to capture the distinguishing feature capability of the teacher network. Experimental results show that the GIP method outperforms knowledge distillation using GANs as well as training using only knowledge distillation.

Robust knowledge distillation based on feature variance against backdoored teacher model

Identifying lethal vector and non-vector mosquitoes can become difficult for a layperson and sometimes even for experts, considering their visual similarities. Recently, deep learning (DL) became a solution to assist in differentiating the two mosquito types to reduce infections and enhance actions against them. However, the existing methods employed to develop a DL model for such a task tend to require massive amounts of computing resources and steps, making them impractical. Based on existing methods, most researchers rely on training pre-trained state-of-the-art (SOTA) deep convolutional neural networks (DCNN), which usually require about a million parameters to train. Hence, this method proposes an approach to craft a model with a far lower computing cost while attaining similar or even significantly better performance than pre-existing models in automating the taxonomy of several mosquitoes. This method combines the approach of layer-wise compression and feature fusion with enhanced residual learning that consists of a self-normalizing activation and depthwise convolutions.•The proposed method yielded a model that outperformed the most recent and classic state-of-the-art deep convolutional neural network models.•With the help of the modified residual block and knowledge distillation, the proposed method significantly reduced a fused model's cost while maintaining competitive performance.•Unlike other methods, the proposed method had the best performance-to-cost ratio.

A deep collaborative learning approach is introduced in which a chain of randomly wired neural networks is trained simultaneously to improve the overall generalization and form a strong ensemble model. The proposed method takes advantage of functional-preserving transfer learning and knowledge distillation to produce an ensemble model. Knowledge distillation is an effective learning scheme for improving the performance of small neural networks by using the knowledge learned by teacher networks. Most of the previous methods learn from one or more teachers but not in a collaborative way. In this paper, we created a chain of randomly wired neural networks based on a random graph algorithm and collaboratively trained the models using functional-preserving transfer learning, so that the small network in the chain could learn from the largest one simultaneously. The training method applies knowledge distillation between randomly wired models, where each model is considered as a teacher to the next model in the chain. The decision of multiple chains of models can be combined to produce a robust ensemble model. The proposed method is evaluated on CIFAR-10, CIFAR-100, and TinyImageNet. The experimental results show that the collaborative training significantly improved the generalization of each model, which allowed for obtaining a small model that can mimic the performance of a large model and produce a more robust ensemble approach.

We discuss methods for visualizing neural network decision boundaries and decision regions. We use these visual-izations to investigate issues related to reproducibility and generalization in neural network training. We observe that changes in model architecture (and its associate inductive bias) cause visible changes in decision boundaries, while multiple runs with the same architecture yield results with strong similarities, especially in the case of wide architectures. We also use decision boundary methods to visualize double descent phenomena. We see that decision boundary reproducibility depends strongly on model width. Near the threshold of interpolation, neural network decision bound-aries become fragmented into many small decision regions, and these regions are non-reproducible. Meanwhile, very narrows and very wide networks have high levels of re-producibility in their decision boundaries with relatively few decision regions. We discuss how our observations re-late to the theory of double descent phenomena in convex models. Code is available at https://github.com/somepago/dbViz.

Disentangling the intrinsic feature from the related feature in image classification using knowledge distillation and object replacement

This paper presents a visual sentiment prediction method using cross-way few-shot learning based on knowledge distillation. Previous studies on visual sentiment prediction methods have focused only on one sentiment dataset although there are several sentiment datasets following different sentiment theories. Originally, sentiments are abstract notions common to humans regardless of the difference between sentiment theories. Thus, the use of knowledge obtained from several sentiment datasets can be the effective way to realize robust visual sentiment prediction. To collaboratively use sentiment datasets, there are the following two concerns: training of the model introducing different sentiment theories and prediction of a newly given sample whose sentiment theory is unknown. Thus, we focus on knowledge distillation, which can improve the generalization ability of several tasks, and effective training becomes feasible. In addition, to deal with the different numbers of sentiment labels in a test phase, we newly introduce the cross-way few-shot learning scheme into knowledge distillation. The main contribution in this paper is to integrate knowledge distillation and the cross-way approach for the visual sentiment prediction, and this is the first work for dealing with the difference of sentiment datasets used in the training and test phases. In the end of this paper, the effectiveness of the proposed method is confirmed through experiments using several open datasets.

Knowledge distillation (KD) is an effective model compression technique where a compact student network is taught to mimic the behavior of a complex and highly trained teacher network. In contrast, Mutual Learning (ML) provides an alternative strategy where multiple simple student networks benefit from sharing knowledge, even in the absence of a powerful but static teacher network. Motivated by these findings, we propose a single-teacher, multi-student framework that leverages both KD and ML to achieve better performance. Furthermore, an online distillation strategy is utilized to train the teacher and students simultaneously. To evaluate the performance of the proposed approach, extensive experiments were conducted using three different versions of teacher-student networks on benchmark biomedical classification (MSI vs. MSS) and object detection (Polyp Detection) tasks. Ensemble of student networks trained in the proposed manner achieved better results than the ensemble of students trained using KD or ML individually, establishing the benefit of augmenting knowledge transfer from teacher to students with peer-to-peer learning between students.

Knowledge distillation aims to distill knowledge from teacher networks to train student networks. Distilling intermediate features has attracted much attention in recent years as it can be flexibly applied in various fields such as image classification, object detection and semantic segmentation. A critical obstacle of feature-based knowledge distillation is the dimension gap between the intermediate features of teacher and student, and plenty of methods have been proposed to resolve this problem. However, these works usually implement feature uniformization in an unsupervised way, lacking guidance to help the student network learn meaningful mapping functions in the uniformization process. Moreover, the dimension uniformization process of the student and teacher network is usually not equivalent as the mapping functions are different. To this end, some factors of the feature are discarded during parametric feature alignment, or some factors are blended in some non-parametric operations. In this paper, we propose a novel semantic-aware knowledge distillation scheme to solve these problems. We build a standalone feature-based classification branch to extract semantic-aware knowledge for better guiding the learning process of the student network. In addition, to avoid the inequivalence of feature uniformization between teacher and student, we design a novel parameter-free self-attention operation that can convert features of different dimensions into vectors of the same length. Experimental results show that the proposed knowledge distillation scheme outperforms existing feature-based distillation methods on the widely used CIFAR-100 and CINIC-10 datasets.

Open World Object Detection (OWOD) is a novel computer vision task with a considerable challenge, bridging the gap between classic object detection (OD) and real-world object detection. In addition to detecting and classifying seen/known objects, OWOD algorithms are expected to localize all potential unseen/unknown objects and incrementally learn them. The large pre-trained vision-language grounding models (VLM, e.g., GLIP) have rich knowledge about the open world, but are limited by text prompts and cannot localize indescribable objects. However, there are many detection scenarios in which pre-defined language descriptions are unavailable during inference. In this paper, we attempt to specialize the VLM model for OWOD tasks by distilling its open-world knowledge into a language-agnostic detector. Surprisingly, we observe that the simple knowledge distillation approach leads to unexpected performance for unknown object detection, even with a small amount of data. Unfortunately, knowledge distillation for unknown objects severely affects the learning of detectors with conventional structures, leading to catastrophic damage to the model's ability to learn about known objects. To alleviate these problems, we propose the down-weight training strategy for knowledge distillation from vision-language model to single visual modality one. Meanwhile, we propose the cascade decoupled decoders that decouple the learning of localization and recognition to reduce the impact of category interactions of known and unknown objects on the localization learning process. Ablation experiments demonstrate that both of them are effective in mitigating the impact of open-world knowledge distillation on the learning of known objects. Additionally, to alleviate the current lack of comprehensive benchmarks for evaluating the ability of the open-world detector to detect unknown objects in the open world, we refine the benchmark for evaluating the performance of unknown object detection by augmenting annotations for unknown objects which we name"IntensiveSet$\scriptstyle\spadesuit$♠". Comprehensive experiments performed on OWOD, MS-COCO, and our proposed benchmarks demonstrate the effectiveness of our methods.

Encoding facial expressions via action units (AUs) has been found to be effective in resolving the ambiguity issue among different expressions. Therefore, AU detection plays an important role for emotion analysis. While a number of AU detection methods have been proposed for common facial expressions, there is very limited study for micro-expression AU detection. Micro-expression AU detection is challenging because of the weakness of micro-expression appearance and the spontaneous characteristic leading to difficult collection, thus has small-scale datasets. In this paper, we focus on the micro-expression AU detection and expect to contribute to the community. To address above issues, a novel dual-view attentive similarity-preserving distillation method is proposed for robust micro-expression AU detection by leveraging massive facial expressions in the wild. Through such an attentive similarity-preserving distillation method, we break the domain shift problem and essential AU knowledge from common facial AUs is efficiently distilled. Furthermore, considering that the generalization ability of teacher network is important for knowledge distillation, a semi-supervised co-training approach is developed to construct a generalized teacher network for learning discriminative AU representation. Extensive experiments have demonstrated that our proposed knowledge distillation method can effectively distill and transfer the cross-domain knowledge for robust micro-expression AU detection.

Deep Neural Networks are often vulnerable to adversarial attacks. Neural Architecture Search (NAS), one of the tools for developing novel deep neural architectures, demonstrates superior performance in prediction accuracy in various machine learning applications. However, the performance of a neural architecture discovered by NAS against adversarial attacks has not been sufficiently studied, especially under the regime of knowledge distillation. Given the presence of a robust teacher, we investigate if NAS would produce a robust neural architecture by inheriting robustness from the teacher. In this paper, we propose Robust Neural Architecture Search by Cross-Layer knowledge distillation (RNAS-CL), a novel NAS algorithm that improves the robustness of NAS by learning from a robust teacher through cross-layer knowledge distillation. Unlike previous knowledge distillation methods that encourage close student-teacher output only in the last layer, RNAS-CL automatically searches for the best teacher layer to supervise each student layer. Experimental results demonstrate the effectiveness of RNAS-CL and show that RNAS-CL produces compact and adversarially robust neural architectures. Our results point to new approaches for finding compact and robust neural architecture for many applications. The code of RNAS-CL is available at https://github.com/Statistical-Deep-Learning/RNAS-CL.