PyTorchKR Papers

최신 논문

총 4,526편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2205.02850

A Deep Reinforcement Learning Framework for Rapid Diagnosis of Whole Slide Pathological Images

The deep neural network is a research hotspot for histopathological image analysis, which can improve the efficiency and accuracy of diagnosis for pathologists or be used for disease screening. The whole slide pathological image can reach one gigapixel and contains abundant tissue feature information, which needs to be divided into a lot of patches in the training and inference stages. This will lead to a long convergence time and large memory consumption. Furthermore, well-annotated data sets are also in short supply in the field of digital pathology. Inspired by the pathologist's clinical diagnosis process, we propose a weakly supervised deep reinforcement learning framework, which can greatly reduce the time required for network inference. We use neural network to construct the search model and decision model of reinforcement learning agent respectively. The search model predicts the next action through the image features of different magnifications in the current field of view, and the decision model is used to return the predicted probability of the current field of view image. In addition, an expert-guided model is constructed by multi-instance learning, which not only provides rewards for search model, but also guides decision model learning by the knowledge distillation method. Experimental results show that our proposed method can achieve fast inference and accurate prediction of whole slide images without any pixel-level annotations.

arXiv:2205.03273공식 코드 · Papers with Code

Collective Relevance Labeling for Passage Retrieval

Deep learning for Information Retrieval (IR) requires a large amount of high-quality query-document relevance labels, but such labels are inherently sparse. Label smoothing redistributes some observed probability mass over unobserved instances, often uniformly, uninformed of the true distribution. In contrast, we propose knowledge distillation for informed labeling, without incurring high computation overheads at evaluation time. Our contribution is designing a simple but efficient teacher model which utilizes collective knowledge, to outperform state-of-the-arts distilled from a more complex teacher model. Specifically, we train up to x8 faster than the state-of-the-art teacher, while distilling the rankings better. Our code is publicly available at https://github.com/jihyukkim-nlp/CollectiveKD

arXiv:2205.02340공식 코드 · Papers with Code

Knowledge Distillation of Russian Language Models with Reduction of Vocabulary

Today, transformer language models serve as a core component for majority of natural language processing tasks. Industrial application of such models requires minimization of computation time and memory footprint. Knowledge distillation is one of approaches to address this goal. Existing methods in this field are mainly focused on reducing the number of layers or dimension of embeddings/hidden representations. Alternative option is to reduce the number of tokens in vocabulary and therefore the embeddings matrix of the student model. The main problem with vocabulary minimization is mismatch between input sequences and output class distributions of a teacher and a student models. As a result, it is impossible to directly apply KL-based knowledge distillation. We propose two simple yet effective alignment techniques to make knowledge distillation to the students with reduced vocabulary. Evaluation of distilled models on a number of common benchmarks for Russian such as Russian SuperGLUE, SberQuAD, RuSentiment, ParaPhaser, Collection-3 demonstrated that our techniques allow to achieve compression from $17\times$ to $49\times$, while maintaining quality of $1.7\times$ compressed student with the full-sized vocabulary, but reduced number of Transformer layers only. We make our code and distilled models available.

arXiv:2205.02604

Holistic Approach to Measure Sample-level Adversarial Vulnerability and its Utility in Building Trustworthy Systems

Adversarial attack perturbs an image with an imperceptible noise, leading to incorrect model prediction. Recently, a few works showed inherent bias associated with such attack (robustness bias), where certain subgroups in a dataset (e.g. based on class, gender, etc.) are less robust than others. This bias not only persists even after adversarial training, but often results in severe performance discrepancies across these subgroups. Existing works characterize the subgroup's robustness bias by only checking individual sample's proximity to the decision boundary. In this work, we argue that this measure alone is not sufficient and validate our argument via extensive experimental analysis. It has been observed that adversarial attacks often corrupt the high-frequency components of the input image. We, therefore, propose a holistic approach for quantifying adversarial vulnerability of a sample by combining these different perspectives, i.e., degree of model's reliance on high-frequency features and the (conventional) sample-distance to the decision boundary. We demonstrate that by reliably estimating adversarial vulnerability at the sample level using the proposed holistic metric, it is possible to develop a trustworthy system where humans can be alerted about the incoming samples that are highly likely to be misclassified at test time. This is achieved with better precision when our holistic metric is used over individual measures. To further corroborate the utility of the proposed holistic approach, we perform knowledge distillation in a limited-sample setting. We observe that the student network trained with the subset of samples selected using our combined metric performs better than both the competing baselines, viz., where samples are selected randomly or based on their distances to the decision boundary.

arXiv:2205.02468공식 코드 · Papers with Code

Alignahead: Online Cross-Layer Knowledge Extraction on Graph Neural Networks

Existing knowledge distillation methods on graph neural networks (GNNs) are almost offline, where the student model extracts knowledge from a powerful teacher model to improve its performance. However, a pre-trained teacher model is not always accessible due to training cost, privacy, etc. In this paper, we propose a novel online knowledge distillation framework to resolve this problem. Specifically, each student GNN model learns the extracted local structure from another simultaneously trained counterpart in an alternating training procedure. We further develop a cross-layer distillation strategy by aligning ahead one student layer with the layer in different depth of another student model, which theoretically makes the structure information spread over all layers. Experimental results on five datasets including PPI, Coauthor-CS/Physics and Amazon-Computer/Photo demonstrate that the student performance is consistently boosted in our collaborative training framework without the supervision of a pre-trained teacher model. In addition, we also find that our alignahead technique can accelerate the model convergence speed and its effectiveness can be generally improved by increasing the student numbers in training. Code is available: https://github.com/GuoJY-eatsTG/Alignahead

arXiv:2205.02399공식 코드 · Papers with Code

Spot-adaptive Knowledge Distillation

Knowledge distillation (KD) has become a well established paradigm for compressing deep neural networks. The typical way of conducting knowledge distillation is to train the student network under the supervision of the teacher network to harness the knowledge at one or multiple spots (i.e., layers) in the teacher network. The distillation spots, once specified, will not change for all the training samples, throughout the whole distillation process. In this work, we argue that distillation spots should be adaptive to training samples and distillation epochs. We thus propose a new distillation strategy, termed spot-adaptive KD (SAKD), to adaptively determine the distillation spots in the teacher network per sample, at every training iteration during the whole distillation period. As SAKD actually focuses on "where to distill" instead of "what to distill" that is widely investigated by most existing works, it can be seamlessly integrated into existing distillation methods to further improve their performance. Extensive experiments with 10 state-of-the-art distillers are conducted to demonstrate the effectiveness of SAKD for improving their distillation performance, under both homogeneous and heterogeneous distillation settings. Code is available at https://github.com/zju-vipa/spot-adaptive-pytorch

arXiv:2205.02359

FedSPLIT: One-Shot Federated Recommendation System Based on Non-negative Joint Matrix Factorization and Knowledge Distillation

Non-negative matrix factorization (NMF) with missing-value completion is a well-known effective Collaborative Filtering (CF) method used to provide personalized user recommendations. However, traditional CF relies on the privacy-invasive collection of users' explicit and implicit feedback to build a central recommender model. One-shot federated learning has recently emerged as a method to mitigate the privacy problem while addressing the traditional communication bottleneck of federated learning. In this paper, we present the first unsupervised one-shot federated CF implementation, named FedSPLIT, based on NMF joint factorization. In our solution, the clients first apply local CF in-parallel to build distinct client-specific recommenders. Then, the privacy-preserving local item patterns and biases from each client are shared with the processor to perform joint factorization in order to extract the global item patterns. Extracted patterns are then aggregated to each client to build the local models via knowledge distillation. In our experiments, we demonstrate the feasibility of our approach with standard recommendation datasets. FedSPLIT can obtain similar results than the state of the art (and even outperform it in certain situations) with a substantial decrease in the number of communications.

arXiv:2205.01997

Attention-based Knowledge Distillation in Multi-attention Tasks: The Impact of a DCT-driven Loss

Knowledge Distillation (KD) is a strategy for the definition of a set of transferability gangways to improve the efficiency of Convolutional Neural Networks. Feature-based Knowledge Distillation is a subfield of KD that relies on intermediate network representations, either unaltered or depth-reduced via maximum activation maps, as the source knowledge. In this paper, we propose and analyse the use of a 2D frequency transform of the activation maps before transferring them. We pose that\textemdash by using global image cues rather than pixel estimates, this strategy enhances knowledge transferability in tasks such as scene recognition, defined by strong spatial and contextual relationships between multiple and varied concepts. To validate the proposed method, an extensive evaluation of the state-of-the-art in scene recognition is presented. Experimental results provide strong evidences that the proposed strategy enables the student network to better focus on the relevant image areas learnt by the teacher network, hence leading to better descriptive features and higher transferred performance than every other state-of-the-art alternative. We publicly release the training and evaluation framework used along this paper at http://www-vpu.eps.uam.es/publications/DCTBasedKDForSceneRecognition.

arXiv:2205.01915공식 코드 · Papers with Code

Generalized Knowledge Distillation via Relationship Matching

The knowledge of a well-trained deep neural network (a.k.a. the "teacher") is valuable for learning similar tasks. Knowledge distillation extracts knowledge from the teacher and integrates it with the target model (a.k.a. the "student"), which expands the student's knowledge and improves its learning efficacy. Instead of enforcing the teacher to work on the same task as the student, we borrow the knowledge from a teacher trained from a general label space -- in this "Generalized Knowledge Distillation (GKD)", the classes of the teacher and the student may be the same, completely different, or partially overlapped. We claim that the comparison ability between instances acts as an essential factor threading knowledge across tasks, and propose the RElationship FacIlitated Local cLassifiEr Distillation (REFILLED) approach, which decouples the GKD flow of the embedding and the top-layer classifier. In particular, different from reconciling the instance-label confidence between models, REFILLED requires the teacher to reweight the hard tuples pushed forward by the student and then matches the similarity comparison levels between instances. An embedding-induced classifier based on the teacher model supervises the student's classification confidence and adaptively emphasizes the most related supervision from the teacher. REFILLED demonstrates strong discriminative ability when the classes of the teacher vary from the same to a fully non-overlapped set w.r.t. the student. It also achieves state-of-the-art performance on standard knowledge distillation, one-step incremental learning, and few-shot learning tasks.

arXiv:2205.01529공식 코드 · Papers with Code

Masked Generative Distillation

Knowledge distillation has been applied to various tasks successfully. The current distillation algorithm usually improves students' performance by imitating the output of the teacher. This paper shows that teachers can also improve students' representation power by guiding students' feature recovery. From this point of view, we propose Masked Generative Distillation (MGD), which is simple: we mask random pixels of the student's feature and force it to generate the teacher's full feature through a simple block. MGD is a truly general feature-based distillation method, which can be utilized on various tasks, including image classification, object detection, semantic segmentation and instance segmentation. We experiment on different models with extensive datasets and the results show that all the students achieve excellent improvements. Notably, we boost ResNet-18 from 69.90% to 71.69% ImageNet top-1 accuracy, RetinaNet with ResNet-50 backbone from 37.4 to 41.0 Boundingbox mAP, SOLO based on ResNet-50 from 33.1 to 36.2 Mask mAP and DeepLabV3 based on ResNet-18 from 73.20 to 76.02 mIoU. Our codes are available at https://github.com/yzd-v/MGD.

arXiv:2205.00479공식 코드 · Papers with Code

Nearest Neighbor Knowledge Distillation for Neural Machine Translation

k-nearest-neighbor machine translation (NN-MT), proposed by Khandelwal et al. (2021), has achieved many state-of-the-art results in machine translation tasks. Although effective, NN-MT requires conducting NN searches through the large datastore for each decoding step during inference, prohibitively increasing the decoding cost and thus leading to the difficulty for the deployment in real-world applications. In this paper, we propose to move the time-consuming NN search forward to the preprocessing phase, and then introduce Nearest Neighbor Knowledge Distillation (NN-KD) that trains the base NMT model to directly learn the knowledge of NN. Distilling knowledge retrieved by NN can encourage the NMT model to take more reasonable target tokens into consideration, thus addressing the overcorrection problem. Extensive experimental results show that, the proposed method achieves consistent improvement over the state-of-the-art baselines including NN-MT, while maintaining the same training and decoding speed as the standard NMT model.

arXiv:2205.00258공식 코드 · Papers with Code

EasyNLP: A Comprehensive and Easy-to-use Toolkit for Natural Language Processing

The success of Pre-Trained Models (PTMs) has reshaped the development of Natural Language Processing (NLP). Yet, it is not easy to obtain high-performing models and deploy them online for industrial practitioners. To bridge this gap, EasyNLP is designed to make it easy to build NLP applications, which supports a comprehensive suite of NLP algorithms. It further features knowledge-enhanced pre-training, knowledge distillation and few-shot learning functionalities for large-scale PTMs, and provides a unified framework of model training, inference and deployment for real-world applications. Currently, EasyNLP has powered over ten business units within Alibaba Group and is seamlessly integrated to the Platform of AI (PAI) products on Alibaba Cloud. The source code of our EasyNLP toolkit is released at GitHub (https://github.com/alibaba/EasyNLP).

arXiv:2205.00706

FedDKD: Federated Learning with Decentralized Knowledge Distillation

The performance of federated learning in neural networks is generally influenced by the heterogeneity of the data distribution. For a well-performing global model, taking a weighted average of the local models, as done by most existing federated learning algorithms, may not guarantee consistency with local models in the space of neural network maps. In this paper, we propose a novel framework of federated learning equipped with the process of decentralized knowledge distillation (FedDKD) (i.e., without data on the server). The FedDKD introduces a module of decentralized knowledge distillation (DKD) to distill the knowledge of the local models to train the global model by approaching the neural network map average based on the metric of divergence defined in the loss function, other than only averaging parameters as done in literature. Numeric experiments on various heterogeneous datasets reveal that FedDKD outperforms the state-of-the-art methods with more efficient communication and training in a few DKD steps, especially on some extremely heterogeneous datasets.

arXiv:2204.13873

Multiple Degradation and Reconstruction Network for Single Image Denoising via Knowledge Distillation

Single image denoising (SID) has achieved significant breakthroughs with the development of deep learning. However, the proposed methods are often accompanied by plenty of parameters, which greatly limits their application scenarios. Different from previous works that blindly increase the depth of the network, we explore the degradation mechanism of the noisy image and propose a lightweight Multiple Degradation and Reconstruction Network (MDRN) to progressively remove noise. Meanwhile, we propose two novel Heterogeneous Knowledge Distillation Strategies (HMDS) to enable MDRN to learn richer and more accurate features from heterogeneous models, which make it possible to reconstruct higher-quality denoised images under extreme conditions. Extensive experiments show that our MDRN achieves favorable performance against other SID models with fewer parameters. Meanwhile, plenty of ablation studies demonstrate that the introduced HMDS can improve the performance of tiny models or the model under high noise levels, which is extremely useful for related applications.

arXiv:2204.13679공식 코드 · Papers with Code

Curriculum Learning for Dense Retrieval Distillation

Recent work has shown that more effective dense retrieval models can be obtained by distilling ranking knowledge from an existing base re-ranking model. In this paper, we propose a generic curriculum learning based optimization framework called CL-DRD that controls the difficulty level of training data produced by the re-ranking (teacher) model. CL-DRD iteratively optimizes the dense retrieval (student) model by increasing the difficulty of the knowledge distillation data made available to it. In more detail, we initially provide the student model coarse-grained preference pairs between documents in the teacher's ranking and progressively move towards finer-grained pairwise document ordering requirements. In our experiments, we apply a simple implementation of the CL-DRD framework to enhance two state-of-the-art dense retrieval models. Experiments on three public passage retrieval datasets demonstrate the effectiveness of our proposed framework.

arXiv:2204.12915

Improving Feature Generalizability with Multitask Learning in Class Incremental Learning

Many deep learning applications, like keyword spotting, require the incorporation of new concepts (classes) over time, referred to as Class Incremental Learning (CIL). The major challenge in CIL is catastrophic forgetting, i.e., preserving as much of the old knowledge as possible while learning new tasks. Various techniques, such as regularization, knowledge distillation, and the use of exemplars, have been proposed to resolve this issue. However, prior works primarily focus on the incremental learning step, while ignoring the optimization during the base model training. We hypothesize that a more transferable and generalizable feature representation from the base model would be beneficial to incremental learning. In this work, we adopt multitask learning during base model training to improve the feature generalizability. Specifically, instead of training a single model with all the base classes, we decompose the base classes into multiple subsets and regard each of them as a task. These tasks are trained concurrently and a shared feature extractor is obtained for incremental learning. We evaluate our approach on two datasets under various configurations. The results show that our approach enhances the average incremental learning accuracy by up to 5.5%, which enables more reliable and accurate keyword spotting over time. Moreover, the proposed approach can be combined with many existing techniques and provides additional performance gain.

arXiv:2204.12493공식 코드 · Papers with Code

One-shot Federated Learning without Server-side Training

Federated Learning (FL) has recently made significant progress as a new machine learning paradigm for privacy protection. Due to the high communication cost of traditional FL, one-shot federated learning is gaining popularity as a way to reduce communication cost between clients and the server. Most of the existing one-shot FL methods are based on Knowledge Distillation; however, {distillation based approach requires an extra training phase and depends on publicly available data sets or generated pseudo samples.} In this work, we consider a novel and challenging cross-silo setting: performing a single round of parameter aggregation on the local models without server-side training. In this setting, we propose an effective algorithm for Model Aggregation via Exploring Common Harmonized Optima (MA-Echo), which iteratively updates the parameters of all local models to bring them close to a common low-loss area on the loss surface, without harming performance on their own data sets at the same time. Compared to the existing methods, MA-Echo can work well even in extremely non-identical data distribution settings where the support categories of each local model have no overlapped labels with those of the others. We conduct extensive experiments on two popular image classification data sets to compare the proposed method with existing methods and demonstrate the effectiveness of MA-Echo, which clearly outperforms the state-of-the-arts. The source code can be accessed in \url{https://github.com/FudanVI/MAEcho}.

arXiv:2204.12997

DearKD: Data-Efficient Early Knowledge Distillation for Vision Transformers

Transformers are successfully applied to computer vision due to their powerful modeling capacity with self-attention. However, the excellent performance of transformers heavily depends on enormous training images. Thus, a data-efficient transformer solution is urgently needed. In this work, we propose an early knowledge distillation framework, which is termed as DearKD, to improve the data efficiency required by transformers. Our DearKD is a two-stage framework that first distills the inductive biases from the early intermediate layers of a CNN and then gives the transformer full play by training without distillation. Further, our DearKD can be readily applied to the extreme data-free case where no real images are available. In this case, we propose a boundary-preserving intra-divergence loss based on DeepInversion to further close the performance gap against the full-data counterpart. Extensive experiments on ImageNet, partial ImageNet, data-free setting and other downstream tasks prove the superiority of DearKD over its baselines and state-of-the-art methods.

arXiv:2204.12833

Transfer Learning with Pre-trained Conditional Generative Models

Transfer learning is crucial in training deep neural networks on new target tasks. Current transfer learning methods always assume at least one of (i) source and target task label spaces overlap, (ii) source datasets are available, and (iii) target network architectures are consistent with source ones. However, holding these assumptions is difficult in practical settings because the target task rarely has the same labels as the source task, the source dataset access is restricted due to storage costs and privacy, and the target architecture is often specialized to each task. To transfer source knowledge without these assumptions, we propose a transfer learning method that uses deep generative models and is composed of the following two stages: pseudo pre-training (PP) and pseudo semi-supervised learning (P-SSL). PP trains a target architecture with an artificial dataset synthesized by using conditional source generative models. P-SSL applies SSL algorithms to labeled target data and unlabeled pseudo samples, which are generated by cascading the source classifier and generative models to condition them with target samples. Our experimental results indicate that our method can outperform the baselines of scratch training and knowledge distillation.

arXiv:2204.12819공식 코드 · Papers with Code

Conformer and Blind Noisy Students for Improved Image Quality Assessment

Generative models for image restoration, enhancement, and generation have significantly improved the quality of the generated images. Surprisingly, these models produce more pleasant images to the human eye than other methods, yet, they may get a lower perceptual quality score using traditional perceptual quality metrics such as PSNR or SSIM. Therefore, it is necessary to develop a quantitative metric to reflect the performance of new algorithms, which should be well-aligned with the person's mean opinion score (MOS). Learning-based approaches for perceptual image quality assessment (IQA) usually require both the distorted and reference image for measuring the perceptual quality accurately. However, commonly only the distorted or generated image is available. In this work, we explore the performance of transformer-based full-reference IQA models. We also propose a method for IQA based on semi-supervised knowledge distillation from full-reference teacher models into blind student models using noisy pseudo-labeled data. Our approaches achieved competitive results on the NTIRE 2022 Perceptual Image Quality Assessment Challenge: our full-reference model was ranked 4th, and our blind noisy student was ranked 3rd among 70 participants, each in their respective track.