Revisiting Distillation and Incremental Classifier Learning
National University of Sciences and Technology（国家科技大学）；School of Electrical Engineering and Computer Science（电子工程与计算机科学学院）；National Center of Artificial Intelligence（国家人工智能中心）
One of the key differences between the learning mechanism of humans and Artificial Neural Networks (ANNs) is the ability of humans to learn one task at a time. ANNs, on the other hand, can only learn multiple tasks simultaneously. Any attempts at learning new tasks incrementally cause them to completely forget about previous tasks. This lack of ability to learn incrementally, called Catastrophic Forgetting, is considered a major hurdle in building a true AI system. In this paper, our goal is to isolate the truly effective existing ideas for incremental learning from those that only work under certain conditions. To this end, we first thoroughly analyze the current state of the art (iCaRL) method for incremental learning and demonstrate that the good performance of the system is not because of the reasons presented in the existing literature. We conclude that the success of iCaRL is primarily due to knowledge distillation and recognize a key limitation of knowledge distillation, i.e, it often leads to bias in classifiers. Finally, we propose a dynamic threshold moving algorithm that is able to successfully remove this bias. We demonstrate the effectiveness of our algorithm on CIFAR100 and MNIST datasets showing near-optimal results. Our implementation is available at https://github.com/Khurramjaved96/incremental-learning.
Knowledge distillation is the procedure of transferring "knowledge" from a large model (the teacher) to a more compact one (the student), often being used in the context of model compression. When both models have the same architecture, this procedure is called self-distillation. Several works have anecdotally shown that a self-distilled student can outperform the teacher on held-out data. In this work, we systematically study self-distillation in a number of settings. We first show that even with a highly accurate teacher, self-distillation allows a student to surpass the teacher in all cases…
Knowledge Distillation (KD) has recently emerged as a popular method for compressing neural networks. In recent studies, generalized distillation methods that find parameters and architectures of student models at the same time have been proposed. Still, this search method requires a lot of computation to search for architectures and has the disadvantage of considering only convolutional blocks in their search space. This paper introduces a new algorithm, coined as Trust Region Aware architecture search to Distill knowledge Effectively (TRADE), that rapidly finds effective student architecture…
The existing solutions for object detection distillation rely on the availability of both a teacher model and ground-truth labels. We propose a new perspective to relax this constraint. In our framework, a student is first trained with pseudo labels generated by the teacher, and then fine-tuned using labeled data, if any available. Extensive experiments demonstrate improvements over existing object detection distillation algorithms. In addition, decoupling the teacher and ground-truth distillation in this framework provides interesting properties such: as 1) using unlabeled data to further imp…
University of Maryland（马里兰大学）；Adobe Research（奥多比研究中心）
Incremental learning targets at achieving good performance on new categories without forgetting old ones. Knowledge distillation has been shown critical in preserving the performance on old classes. Conventional methods, however, sequentially distill knowledge only from the last model, leading to performance degradation on the old classes in later incremental learning steps. In this paper, we propose a multi-model and multi-level knowledge distillation strategy. Instead of sequentially distilling knowledge only from the last model, we directly leverage all previous model snapshots. In addition…
Deep learning architectures have shown remarkable results in scene understanding problems, however they exhibit a critical drop of performances when they are required to learn incrementally new tasks without forgetting old ones. This catastrophic forgetting phenomenon impacts on the deployment of artificial intelligence in real world scenarios where systems need to learn new and different representations over time. Current approaches for incremental learning deal only with image classification and object detection tasks, while in this work we formally introduce incremental learning for semanti…
Singapore University of Technology and Design (SUTD)（新加坡科技设计大学）
This work investigates the compatibility between label smoothing (LS) and knowledge distillation (KD). Contemporary findings addressing this thesis statement take dichotomous standpoints: Muller et al. (2019) and Shen et al. (2021b). Critically, there is no effort to understand and resolve these contradictory findings, leaving the primal question -- to smooth or not to smooth a teacher network? -- unanswered. The main contributions of our work are the discovery, analysis and validation of systematic diffusion as the missing concept which is instrumental in understanding and resolving these con…
College of Computer Science, Nankai University（南开大学计算机学院）；MICC, University of Florence（佛罗伦萨大学MICC）；Huawei Kirin Solution（华为麒麟解决方案）
Most meta-learning approaches assume the existence of a very large set of labeled data available for episodic meta-learning of base knowledge. This contrasts with the more realistic continual learning paradigm in which data arrives incrementally in the form of tasks containing disjoint classes. In this paper we consider this problem of Incremental Meta-Learning (IML) in which classes are presented incrementally in discrete tasks. We propose an approach to IML, which we call Episodic Replay Distillation (ERD), that mixes classes from the current task with class exemplars from previous tasks whe…
University of Toronto（多伦多大学）；LG AI Research（LG人工智能研究院）
Online class-incremental continual learning (CL) studies the problem of learning new classes continually from an online non-stationary data stream, intending to adapt to new data while mitigating catastrophic forgetting. While memory replay has shown promising results, the recency bias in online learning caused by the commonly used Softmax classifier remains an unsolved challenge. Although the Nearest-Class-Mean (NCM) classifier is significantly undervalued in the CL community, we demonstrate that it is a simple yet effective substitute for the Softmax classifier. It addresses the recency bias…
Australian National University（澳大利亚国立大学）；Data61-CSIRO Australia（澳大利亚联邦科学与工业研究组织Data61）；North South University Bangladesh（孟加拉国南北大学）；Monash University Australia（澳大利亚莫纳什大学）
Few-shot class incremental learning (FSCIL) portrays the problem of learning new concepts gradually, where only a few examples per concept are available to the learner. Due to the limited number of examples for training, the techniques developed for standard incremental learning cannot be applied verbatim to FSCIL. In this work, we introduce a distillation algorithm to address the problem of FSCIL and propose to make use of semantic information during training. To this end, we make use of word embeddings as semantic information which is cheap to obtain and which facilitate the distillation pro…
Heuritech；Sorbonne University（巴黎索邦大学）；University of Campinas（坎皮纳斯大学）
Lifelong learning has attracted much attention, but existing works still struggle to fight catastrophic forgetting and accumulate knowledge over long stretches of incremental learning. In this work, we propose PODNet, a model inspired by representation learning. By carefully balancing the compromise between remembering the old classes and learning new ones, PODNet fights catastrophic forgetting, even over very long runs of small incremental tasks --a setting so far unexplored by current works. PODNet innovates on existing art with an efficient spatial-based distillation-loss applied throughout…
Quantum-Gated Task-interaction Knowledge Distillation for Pre-trained Model-based Class-Incremental Learning
School of Information and Communication Engineering, BUPT（北京邮电大学信息与通信工程学院）；BUPT: Beijing University of Posts and Telecommunications（北京邮电大学）；Engineering Research Center for Information Network, Ministry of Education（教育部信息网络工程研究中心）
Knowledge distillation (KD) is a highly promising method for mitigating the computational problems of pre-trained language models (PLMs). Among various KD approaches, Intermediate Layer Distillation (ILD) has been a de facto standard KD method with its performance efficacy in the NLP field. In this paper, we find that existing ILD methods are prone to overfitting to training datasets, although these methods transfer more information than the original KD. Next, we present the simple observations to mitigate the overfitting of ILD: distilling only the last Transformer layer and conducting ILD on…
University of Science and Technology of China（中国科学技术大学）；Alibaba Group（阿里巴巴集团）；University of Southern California（南加州大学）；University of Sydney（悉尼大学）
Knowledge Distillation (KD) is a popular technique to transfer knowledge from a teacher model or ensemble to a student model. Its success is generally attributed to the privileged information on similarities/consistency between the class distributions or intermediate feature representations of the teacher model and the student model. However, directly pushing the student model to mimic the probabilities/features of the teacher model to a large extent limits the student model in learning undiscovered knowledge/features. In this paper, we propose a novel inheritance and exploration knowledge dis…
Face to Cartoon Incremental Super-Resolution using Knowledge Distillation
Indian Institute of Information Technology, Allahabad（印度阿拉哈巴德信息技术学院）；Sorbonne Centre for Artificial Intelligence, Sorbonne University Abu Dhabi（索邦大学阿布扎比分校索邦人工智能中心）
On Distilling the Displacement Knowledge for Few-Shot Class-Incremental Learning
School of Computer Science and Engineering, Southeast University（东南大学计算机科学与工程学院）；Key Laboratory of New Generation Artificial Intelligence Technology and Its Interdisciplinary Applications (Southeast University), Ministry of Education（教育部新一代人工智能技术及交叉应用重点实验室（东南大学））
Huazhong University of Science and Technology（华中科技大学）
Surface defect detection is one of the most essential processes for industrial quality inspection. Deep learning-based surface defect detection methods have shown great potential. However, the well-performed models usually require large training data and can only detect defects that appeared in the training stage. When facing incremental few-shot data, defect detection models inevitably suffer from catastrophic forgetting and misclassification problem. To solve these problems, this paper proposes a new knowledge distillation network, called Dual Knowledge Align Network (DKAN). The proposed DKA…
Incremental learning methods can learn new classes continually by distilling knowledge from the last model (as a teacher model) to the current model (as a student model) in the sequentially learning process. However, these methods cannot work for Incremental Implicitly-Refined Classification (IIRC), an incremental learning extension where the incoming classes could have two granularity levels, a superclass label and a subclass label. This is because the previously learned superclass knowledge may be occupied by the subclass knowledge learned sequentially. To solve this problem, we propose a no…
CD^2: Constrained Dataset Distillation for Few-Shot Class-Incremental Learning
Institute of Information Engineering, Chinese Academy of Sciences（中国科学院信息工程研究所）；School of Cyber Security, University of Chinese Academy of Sciences（中国科学院大学网络安全学院）；Hong Kong University of Science and Technology (Guangzhou)（香港科技大学（广州））
We present a novel class incremental learning approach based on deep neural networks, which continually learns new tasks with limited memory for storing examples in the previous tasks. Our algorithm is based on knowledge distillation and provides a principled way to maintain the representations of old models while adjusting to new tasks effectively. The proposed method estimates the relationship between the representation changes and the resulting loss increases incurred by model updates. It minimizes the upper bound of the loss increases using the representations, which exploits the estimated…
Zhejiang University（浙江大学）；University of Illinois Urbana–Champaign（伊利诺伊大学厄巴纳-香槟分校）；University of Science and Technology Beijing（北京科技大学）
Intelligent fault diagnosis has made extraordinary advancements currently. Nonetheless, few works tackle class-incremental learning for fault diagnosis under limited fault data, i.e., imbalanced and long-tailed fault diagnosis, which brings about various notable challenges. Initially, it is difficult to extract discriminative features from limited fault data. Moreover, a well-trained model must be retrained from scratch to classify the samples from new classes, thus causing a high computational burden and time consumption. Furthermore, the model may suffer from catastrophic forgetting when tra…
Multi-Granularity Class Prototype Topology Distillation for Class-Incremental Source-Free Unsupervised Domain Adaptation
Hangzhou Dianzi University（杭州电子科技大学）；Macao Polytechnic University（澳门理工大学）；Beijing Institute of Technology（北京理工大学）；Institute of Computing Technology, Chinese Academy of Sciences（中国科学院计算技术研究所）
National University of Defense Technology（国防科技大学）；University of Oulu（奥卢大学）
Given a model well-trained with a large-scale base dataset, Few-Shot Class-Incremental Learning (FSCIL) aims at incrementally learning novel classes from a few labeled samples by avoiding overfitting, without catastrophically forgetting all encountered classes previously. Currently, semi-supervised learning technique that harnesses freely-available unlabeled data to compensate for limited labeled data can boost the performance in numerous vision tasks, which heuristically can be applied to tackle issues in FSCIL, i.e., the Semi-supervised FSCIL (Semi-FSCIL). So far, very limited work focuses o…
National University of Singapore（新加坡国立大学）；Huawei Noah’s Ark Lab（华为诺亚方舟实验室）
Knowledge Distillation (KD) aims to distill the knowledge of a cumbersome teacher model into a lightweight student model. Its success is generally attributed to the privileged information on similarities among categories provided by the teacher model, and in this sense, only strong teacher models are deployed to teach weaker students in practice. In this work, we challenge this common belief by following experimental observations: 1) beyond the acknowledgment that the teacher can improve the student, the student can also enhance the teacher significantly by reversing the KD procedure; 2) a poo…
The Australian National University（澳大利亚国立大学）；Monash University（莫纳什大学）；Data61-CSIRO（Data61-澳大利亚联邦科学与工业研究组织）
Neural networks notoriously suffer from the problem of catastrophic forgetting, the phenomenon of forgetting the past knowledge when acquiring new knowledge. Overcoming catastrophic forgetting is of significant importance to emulate the process of "incremental learning", where the model is capable of learning from sequential experience in an efficient and robust way. State-of-the-art techniques for incremental learning make use of knowledge distillation towards preventing catastrophic forgetting. Therein, one updates the network while ensuring that the network's responses to previously seen co…
Class Incremental Fault Diagnosis under Limited Fault Data via Supervised Contrastive Knowledge Distillation
ZJU-UIUC Joint Institute at Zhejiang University（浙江大学-伊利诺伊大学厄巴纳-香槟分校联合研究所）；ZJU-UoE Institute at Zhejiang University（浙江大学-爱丁堡大学联合研究所）；College of Biomedical Engineering and Instrument Science at Zhejiang University（浙江大学生物医学工程与仪器科学学院）；College of Computer Science and Technology in Zhejiang University（浙江大学计算机科学与技术学院）
Dept. Computing Science & Alberta Machine Intelligence Institute (Amii), University of Alberta（计算科学系及阿尔伯塔人工智能研究所（Amii），阿尔伯塔大学）；Canada CIFAR AI Chair, Amii（加拿大CIFAR人工智能主席，Amii）
Yonsei University（延世大学）；Korea Institute of Science and Technology (KIST)（韩国科学技术研究院）
Class-incremental semantic segmentation (CISS) labels each pixel of an image with a corresponding object/stuff class continually. To this end, it is crucial to learn novel classes incrementally without forgetting previously learned knowledge. Current CISS methods typically use a knowledge distillation (KD) technique for preserving classifier logits, or freeze a feature extractor, to avoid the forgetting problem. The strong constraints, however, prevent learning discriminative features for novel classes. We introduce a CISS framework that alleviates the forgetting problem and facilitates learni…
Teacher Agent: A Knowledge Distillation-Free Framework for Rehearsal-based Video Incremental Learning
Nanjing University of Information Science and Technology（南京信息工程大学）；Intellifusion（云天励飞）；Australian Institute for Machine Learning（澳大利亚机器学习研究所）；the University of Adelaide（阿德莱德大学）；University of Electronic Science and Technology of China（电子科技大学）
In this paper, we focus on a new and challenging decentralized machine learning paradigm in which there are continuous inflows of data to be addressed and the data are stored in multiple repositories. We initiate the study of data decentralized class-incremental learning (DCIL) by making the following contributions. Firstly, we formulate the DCIL problem and develop the experimental protocol. Secondly, we introduce a paradigm to create a basic decentralized counterpart of typical (centralized) class-incremental learning approaches, and as a result, establish a benchmark for the DCIL study. Thi…
Rotation Augmented Distillation for Exemplar-Free Class Incremental Learning with Detailed Analysis
School of Artificial Intelligence, Sun Yat-sen University（中山大学人工智能学院）；Guangdong Key Laboratory of Big Data Analysis and Processing（广东省大数据分析与处理重点实验室）；Key Laboratory of Machine Intelligence and Advanced Computing, Ministry of Education（教育部机器智能与先进计算重点实验室）
Hierarchical Augmentation and Distillation for Class Incremental Audio-Visual Video Recognition
School of Information Science and Technology, University of Science and Technology of China（中国科学技术大学信息科学技术学院）；Institute of Automation, Chinese Academy of Sciences（中国科学院自动化研究所）；School of Artificial Intelligence, University of Chinese Academy of Sciences（中国科学院大学人工智能学院）
Xidian University（西安电子科技大学）；Northwestern Polytechnical University（西北工业大学）；Hefei Comprehensive National Science Center（合肥综合性国家科学中心）；Nankai University（南开大学）；City University of Hong Kong（香港城市大学）
Due to the constraints of the imaging device and high cost in operation time, computer tomography (CT) scans are usually acquired with low intra-slice resolution. Improving the intra-slice resolution is beneficial to the disease diagnosis for both human experts and computer-aided systems. To this end, this paper builds a novel medical slice synthesis to increase the between-slice resolution. Considering that the ground-truth intermediate medical slices are always absent in clinical practice, we introduce the incremental cross-view mutual distillation strategy to accomplish this task in the sel…
Revisiting Distillation for Continual Learning on Visual Question Localized-Answering in Robotic Surgery
Wellcome/EPSRC Centre for Interventional and Surgical Sciences (WEISS), University College London（伦敦大学学院惠康/工程与物理科学研究委员会介入与外科科学中心（WEISS））；The Chinese University of Hong Kong (CUHK)（香港中文大学（CUHK））；Shun Hing Institute of Advanced Engineering, CUHK（香港中文大学信兴高等工程研究所）
In recent years, deeper and wider neural networks have shown excellent performance in computer vision tasks, while their enormous amount of parameters results in increased computational cost and overfitting. Several methods have been proposed to compress the size of the networks without reducing network performance. Network pruning can reduce redundant and unnecessary parameters from a network. Knowledge distillation can transfer the knowledge of deeper and wider networks to smaller networks. The performance of the smaller network obtained by these methods is bounded by the predefined network.…
This paper proposes two novel knowledge transfer techniques for class-incremental learning (CIL). First, we propose data-free generative replay (DF-GR) to mitigate catastrophic forgetting in CIL by using synthetic samples from a generative model. In the conventional generative replay, the generative model is pre-trained for old data and shared in extra memory for later incremental learning. In our proposed DF-GR, we train a generative model from scratch without using any training data, based on the pre-trained classification model from the past, so we curtail the cost of sharing pre-trained ge…
Fudan University（复旦大学）；Xiamen University（厦门大学）；The Pennsylvania State University（宾夕法尼亚州立大学）；Chinese Academy of Sciences（中国科学院）
Crowd counting has important applications in public safety and pandemic control. A robust and practical crowd counting system has to be capable of continuously learning with the new incoming domain data in real-world scenarios instead of fitting one domain only. Off-the-shelf methods have some drawbacks when handling multiple domains: (1) the models will achieve limited performance (even drop dramatically) among old domains after training images from new domains due to the discrepancies of intrinsic data distributions from various domains, which is called catastrophic forgetting; (2) the well-…
In this paper, we propose an incremental learning method for end-to-end Automatic Speech Recognition (ASR) which enables an ASR system to perform well on new tasks while maintaining the performance on its originally learned ones. To mitigate catastrophic forgetting during incremental learning, we design a novel explainability-based knowledge distillation for ASR models, which is combined with a response-based knowledge distillation to maintain the original model's predictions and the "reason" for the predictions. Our method works without access to the training data of original tasks, which add…
One of the major limitations of deep learning models is that they face catastrophic forgetting in an incremental learning scenario. There have been several approaches proposed to tackle the problem of incremental learning. Most of these methods are based on knowledge distillation and do not adequately utilize the information provided by older task models, such as uncertainty estimation in predictions. The predictive uncertainty provides the distributional information can be applied to mitigate catastrophic forgetting in a deep learning framework. In the proposed work, we consider a Bayesian fo…
Porsche Engineering Services GmbH（保时捷工程服务有限公司）；Fraunhofer IOSB（弗劳恩霍夫光学、系统技术与图像利用研究所）；Karlsruhe Institute of Technology（卡尔斯鲁厄理工学院）
The field of continual deep learning is an emerging field and a lot of progress has been made. However, concurrently most of the approaches are only tested on the task of image classification, which is not relevant in the field of intelligent vehicles. Only recently approaches for class-incremental semantic segmentation were proposed. However, all of those approaches are based on some form of knowledge distillation. At the moment there are no investigations on replay-based approaches that are commonly used for object recognition in a continual setting. At the same time while unsupervised domai…
