PyTorchKR Papers

최신 논문

총 1,512편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2305.02722공식 코드 · Papers with Code

Avatar Knowledge Distillation: Self-ensemble Teacher Paradigm with Uncertainty

Knowledge distillation is an effective paradigm for boosting the performance of pocket-size model, especially when multiple teacher models are available, the student would break the upper limit again. However, it is not economical to train diverse teacher models for the disposable distillation. In this paper, we introduce a new concept dubbed Avatars for distillation, which are the inference ensemble models derived from the teacher. Concretely, (1) For each iteration of distillation training, various Avatars are generated by a perturbation transformation. We validate that Avatars own higher upper limit of working capacity and teaching ability, aiding the student model in learning diverse and receptive knowledge perspectives from the teacher model. (2) During the distillation, we propose an uncertainty-aware factor from the variance of statistical differences between the vanilla teacher and Avatars, to adjust Avatars' contribution on knowledge transfer adaptively. Avatar Knowledge Distillation AKD is fundamentally different from existing methods and refines with the innovative view of unequal training. Comprehensive experiments demonstrate the effectiveness of our Avatars mechanism, which polishes up the state-of-the-art distillation methods for dense prediction without more extra computational cost. The AKD brings at most 0.7 AP gains on COCO 2017 for Object Detection and 1.83 mIoU gains on Cityscapes for Semantic Segmentation, respectively. Code is available at https://github.com/Gumpest/AvatarKD.

arXiv:2305.01879공식 코드 · Papers with Code

SCOTT: Self-Consistent Chain-of-Thought Distillation

Large language models (LMs) beyond a certain scale, demonstrate the emergent capability of generating free-text rationales for their predictions via chain-of-thought (CoT) prompting. While CoT can yield dramatically improved performance, such gains are only observed for sufficiently large LMs. Even more concerning, there is little guarantee that the generated rationales are consistent with LM's predictions or faithfully justify the decisions. In this work, we propose a faithful knowledge distillation method to learn a small, self-consistent CoT model from a teacher model that is orders of magnitude larger. To form better supervision, we elicit rationales supporting the gold answers from a large LM (teacher) by contrastive decoding, which encourages the teacher to generate tokens that become more plausible only when the answer is considered. To ensure faithful distillation, we use the teacher-generated rationales to learn a student LM with a counterfactual reasoning objective, which prevents the student from ignoring the rationales to make inconsistent predictions. Experiments show that, while yielding comparable end-task performance, our method can generate CoT rationales that are more faithful than baselines do. Further analysis suggests that such a model respects the rationales more when making decisions; thus, we can improve its performance more by refining its rationales.

arXiv:2305.01698공식 코드 · Papers with Code

DeepAqua: Self-Supervised Semantic Segmentation of Wetland Surface Water Extent with SAR Images using Knowledge Distillation

Deep learning and remote sensing techniques have significantly advanced water monitoring abilities; however, the need for annotated data remains a challenge. This is particularly problematic in wetland detection, where water extent varies over time and space, demanding multiple annotations for the same area. In this paper, we present DeepAqua, a self-supervised deep learning model that leverages knowledge distillation (a.k.a. teacher-student model) to eliminate the need for manual annotations during the training phase. We utilize the Normalized Difference Water Index (NDWI) as a teacher model to train a Convolutional Neural Network (CNN) for segmenting water from Synthetic Aperture Radar (SAR) images, and to train the student model, we exploit cases where optical- and radar-based water masks coincide, enabling the detection of both open and vegetated water surfaces. DeepAqua represents a significant advancement in computer vision techniques by effectively training semantic segmentation models without any manually annotated data. Experimental results show that DeepAqua outperforms other unsupervised methods by improving accuracy by 7%, Intersection Over Union by 27%, and F1 score by 14%. This approach offers a practical solution for monitoring wetland water extent changes without needing ground truth data, making it highly adaptable and scalable for wetland conservation efforts.

arXiv:2305.02031공식 코드 · Papers with Code

A Systematic Study of Knowledge Distillation for Natural Language Generation with Pseudo-Target Training

Modern Natural Language Generation (NLG) models come with massive computational and storage requirements. In this work, we study the potential of compressing them, which is crucial for real-world applications serving millions of users. We focus on Knowledge Distillation (KD) techniques, in which a small student model learns to imitate a large teacher model, allowing to transfer knowledge from the teacher to the student. In contrast to much of the previous work, our goal is to optimize the model for a specific NLG task and a specific dataset. Typically in real-world applications, in addition to labeled data there is abundant unlabeled task-specific data, which is crucial for attaining high compression rates via KD. In this work, we conduct a systematic study of task-specific KD techniques for various NLG tasks under realistic assumptions. We discuss the special characteristics of NLG distillation and particularly the exposure bias problem. Following, we derive a family of Pseudo-Target (PT) augmentation methods, substantially extending prior work on sequence-level KD. We propose the Joint-Teaching method, which applies word-level KD to multiple PTs generated by both the teacher and the student. Finally, we validate our findings in an extreme setup with no labeled examples using GPT-4 as the teacher. Our study provides practical model design observations and demonstrates the effectiveness of PT training for task-specific KD in NLG.

arXiv:2305.00620공식 코드 · Papers with Code

Refined Response Distillation for Class-Incremental Player Detection

Detecting players from sports broadcast videos is essential for intelligent event analysis. However, existing methods assume fixed player categories, incapably accommodating the real-world scenarios where categories continue to evolve. Directly fine-tuning these methods on newly emerging categories also exist the catastrophic forgetting due to the non-stationary distribution. Inspired by recent research on incremental object detection (IOD), we propose a Refined Response Distillation (R^2D) method to effectively mitigate catastrophic forgetting for IOD tasks of the players. Firstly, we design a progressive coarse-to-fine distillation region dividing scheme, separating high-value and low-value regions from classification and regression responses for precise and fine-grained regional knowledge distillation. Subsequently, a tailored refined distillation strategy is developed on regions with varying significance to address the performance limitations posed by pronounced feature homogeneity in the IOD tasks of the players. Furthermore, we present the NBA-IOD and Volleyball-IOD datasets as the benchmark and investigate the IOD tasks of the players systematically. Extensive experiments conducted on benchmarks demonstrate that our method achieves state-of-the-art results.The code and datasets are available at https://github.com/beiyan1911/Players-IOD.

arXiv:2304.14668공식 코드 · Papers with Code

Ensemble Modeling with Contrastive Knowledge Distillation for Sequential Recommendation

Sequential recommendation aims to capture users' dynamic interest and predicts the next item of users' preference. Most sequential recommendation methods use a deep neural network as sequence encoder to generate user and item representations. Existing works mainly center upon designing a stronger sequence encoder. However, few attempts have been made with training an ensemble of networks as sequence encoders, which is more powerful than a single network because an ensemble of parallel networks can yield diverse prediction results and hence better accuracy. In this paper, we present Ensemble Modeling with contrastive Knowledge Distillation for sequential recommendation (EMKD). Our framework adopts multiple parallel networks as an ensemble of sequence encoders and recommends items based on the output distributions of all these networks. To facilitate knowledge transfer between parallel networks, we propose a novel contrastive knowledge distillation approach, which performs knowledge transfer from the representation level via Intra-network Contrastive Learning (ICL) and Cross-network Contrastive Learning (CCL), as well as Knowledge Distillation (KD) from the logits level via minimizing the Kullback-Leibler divergence between the output distributions of the teacher network and the student network. To leverage contextual information, we train the primary masked item prediction task alongside the auxiliary attribute prediction task as a multi-task learning scheme. Extensive experiments on public benchmark datasets show that EMKD achieves a significant improvement compared with the state-of-the-art methods. Besides, we demonstrate that our ensemble method is a generalized approach that can also improve the performance of other sequential recommenders. Our code is available at this link: https://github.com/hw-du/EMKD.

arXiv:2304.14800공식 코드 · Papers with Code

Multi-to-Single Knowledge Distillation for Point Cloud Semantic Segmentation

3D point cloud semantic segmentation is one of the fundamental tasks for environmental understanding. Although significant progress has been made in recent years, the performance of classes with few examples or few points is still far from satisfactory. In this paper, we propose a novel multi-to-single knowledge distillation framework for the 3D point cloud semantic segmentation task to boost the performance of those hard classes. Instead of fusing all the points of multi-scans directly, only the instances that belong to the previously defined hard classes are fused. To effectively and sufficiently distill valuable knowledge from multi-scans, we leverage a multilevel distillation framework, i.e., feature representation distillation, logit distillation, and affinity distillation. We further develop a novel instance-aware affinity distillation algorithm for capturing high-level structural knowledge to enhance the distillation efficacy for hard classes. Finally, we conduct experiments on the SemanticKITTI dataset, and the results on both the validation and test sets demonstrate that our method yields substantial improvements compared with the baseline method. The code is available at \Url{https://github.com/skyshoumeng/M2SKD}.

arXiv:2304.13649공식 코드 · Papers with Code

A Symmetric Dual Encoding Dense Retrieval Framework for Knowledge-Intensive Visual Question Answering

Knowledge-Intensive Visual Question Answering (KI-VQA) refers to answering a question about an image whose answer does not lie in the image. This paper presents a new pipeline for KI-VQA tasks, consisting of a retriever and a reader. First, we introduce DEDR, a symmetric dual encoding dense retrieval framework in which documents and queries are encoded into a shared embedding space using uni-modal (textual) and multi-modal encoders. We introduce an iterative knowledge distillation approach that bridges the gap between the representation spaces in these two encoders. Extensive evaluation on two well-established KI-VQA datasets, i.e., OK-VQA and FVQA, suggests that DEDR outperforms state-of-the-art baselines by 11.6% and 30.9% on OK-VQA and FVQA, respectively. Utilizing the passages retrieved by DEDR, we further introduce MM-FiD, an encoder-decoder multi-modal fusion-in-decoder model, for generating a textual answer for KI-VQA tasks. MM-FiD encodes the question, the image, and each retrieved passage separately and uses all passages jointly in its decoder. Compared to competitive baselines in the literature, this approach leads to 5.5% and 8.5% improvements in terms of question answering accuracy on OK-VQA and FVQA, respectively.

arXiv:2304.12777공식 코드 · Papers with Code

Class Attention Transfer Based Knowledge Distillation

Previous knowledge distillation methods have shown their impressive performance on model compression tasks, however, it is hard to explain how the knowledge they transferred helps to improve the performance of the student network. In this work, we focus on proposing a knowledge distillation method that has both high interpretability and competitive performance. We first revisit the structure of mainstream CNN models and reveal that possessing the capacity of identifying class discriminative regions of input is critical for CNN to perform classification. Furthermore, we demonstrate that this capacity can be obtained and enhanced by transferring class activation maps. Based on our findings, we propose class attention transfer based knowledge distillation (CAT-KD). Different from previous KD methods, we explore and present several properties of the knowledge transferred by our method, which not only improve the interpretability of CAT-KD but also contribute to a better understanding of CNN. While having high interpretability, CAT-KD achieves state-of-the-art performance on multiple benchmarks. Code is available at: https://github.com/GzyAftermath/CAT-KD.

arXiv:2304.11923공식 코드 · Papers with Code

Improving Knowledge Distillation via Transferring Learning Ability

Existing knowledge distillation methods generally use a teacher-student approach, where the student network solely learns from a well-trained teacher. However, this approach overlooks the inherent differences in learning abilities between the teacher and student networks, thus causing the capacity-gap problem. To address this limitation, we propose a novel method called SLKD.

arXiv:2304.11393공식 코드 · Papers with Code

Knowledge Distillation from 3D to Bird's-Eye-View for LiDAR Semantic Segmentation

LiDAR point cloud segmentation is one of the most fundamental tasks for autonomous driving scene understanding. However, it is difficult for existing models to achieve both high inference speed and accuracy simultaneously. For example, voxel-based methods perform well in accuracy, while Bird's-Eye-View (BEV)-based methods can achieve real-time inference. To overcome this issue, we develop an effective 3D-to-BEV knowledge distillation method that transfers rich knowledge from 3D voxel-based models to BEV-based models. Our framework mainly consists of two modules: the voxel-to-pillar distillation module and the label-weight distillation module. Voxel-to-pillar distillation distills sparse 3D features to BEV features for middle layers to make the BEV-based model aware of more structural and geometric information. Label-weight distillation helps the model pay more attention to regions with more height information. Finally, we conduct experiments on the SemanticKITTI dataset and Paris-Lille-3D. The results on SemanticKITTI show more than 5% improvement on the test set, especially for classes such as motorcycle and person, with more than 15% improvement. The code can be accessed at https://github.com/fengjiang5/Knowledge-Distillation-from-Cylinder3D-to-PolarNet.

arXiv:2304.11959공식 코드 · Papers with Code

A Forward and Backward Compatible Framework for Few-shot Class-incremental Pill Recognition

Automatic Pill Recognition (APR) systems are crucial for enhancing hospital efficiency, assisting visually impaired individuals, and preventing cross-infection. However, most existing deep learning-based pill recognition systems can only perform classification on classes with sufficient training data. In practice, the high cost of data annotation and the continuous increase in new pill classes necessitate the development of a few-shot class-incremental pill recognition system. This paper introduces the first few-shot class-incremental pill recognition framework, named Discriminative and Bidirectional Compatible Few-Shot Class-Incremental Learning (DBC-FSCIL). It encompasses forward-compatible and backward-compatible learning components. In forward-compatible learning, we propose an innovative virtual class synthesis strategy and a Center-Triplet (CT) loss to enhance discriminative feature learning. These virtual classes serve as placeholders in the feature space for future class updates, providing diverse semantic knowledge for model training. For backward-compatible learning, we develop a strategy to synthesize reliable pseudo-features of old classes using uncertainty quantification, facilitating Data Replay (DR) and Knowledge Distillation (KD). This approach allows for the flexible synthesis of features and effectively reduces additional storage requirements for samples and models. Additionally, we construct a new pill image dataset for FSCIL and assess various mainstream FSCIL methods, establishing new benchmarks. Our experimental results demonstrate that our framework surpasses existing State-of-the-art (SOTA) methods. The code is available at https://github.com/zhang-jinghua/DBC-FSCIL.

arXiv:2304.10668공식 코드 · Papers with Code

Train Your Own GNN Teacher: Graph-Aware Distillation on Textual Graphs

How can we learn effective node representations on textual graphs? Graph Neural Networks (GNNs) that use Language Models (LMs) to encode textual information of graphs achieve state-of-the-art performance in many node classification tasks. Yet, combining GNNs with LMs has not been widely explored for practical deployments due to its scalability issues. In this work, we tackle this challenge by developing a Graph-Aware Distillation framework (GRAD) to encode graph structures into an LM for graph-free, fast inference. Different from conventional knowledge distillation, GRAD jointly optimizes a GNN teacher and a graph-free student over the graph's nodes via a shared LM. This encourages the graph-free student to exploit graph information encoded by the GNN teacher while at the same time, enables the GNN teacher to better leverage textual information from unlabeled nodes. As a result, the teacher and the student models learn from each other to improve their overall performance. Experiments in eight node classification benchmarks in both transductive and inductive settings showcase GRAD's superiority over existing distillation approaches for textual graphs.

arXiv:2304.09388공식 코드 · Papers with Code

An Empirical Study of Leveraging Knowledge Distillation for Compressing Multilingual Neural Machine Translation Models

Knowledge distillation (KD) is a well-known method for compressing neural models. However, works focusing on distilling knowledge from large multilingual neural machine translation (MNMT) models into smaller ones are practically nonexistent, despite the popularity and superiority of MNMT. This paper bridges this gap by presenting an empirical investigation of knowledge distillation for compressing MNMT models. We take Indic to English translation as a case study and demonstrate that commonly used language-agnostic and language-aware KD approaches yield models that are 4-5x smaller but also suffer from performance drops of up to 3.5 BLEU. To mitigate this, we then experiment with design considerations such as shallower versus deeper models, heavy parameter sharing, multi-stage training, and adapters. We observe that deeper compact models tend to be as good as shallower non-compact ones, and that fine-tuning a distilled model on a High-Quality subset slightly boosts translation quality. Overall, we conclude that compressing MNMT models via KD is challenging, indicating immense scope for further research.

arXiv:2304.07969공식 코드 · Papers with Code

Learning to "Segment Anything" in Thermal Infrared Images through Knowledge Distillation with a Large Scale Dataset SATIR

The Segment Anything Model (SAM) is a promptable segmentation model recently introduced by Meta AI that has demonstrated its prowess across various fields beyond just image segmentation. SAM can accurately segment images across diverse fields, and generating various masks. We discovered that this ability of SAM can be leveraged to pretrain models for specific fields. Accordingly, we have proposed a framework that utilizes SAM to generate pseudo labels for pretraining thermal infrared image segmentation tasks. Our proposed framework can effectively improve the accuracy of segmentation results of specific categories beyond the SOTA ImageNet pretrained model. Our framework presents a novel approach to collaborate with models trained with large data like SAM to address problems in special fields. Also, we generated a large scale thermal infrared segmentation dataset used for pretaining, which contains over 100,000 images with pixel-annotation labels. This approach offers an effective solution for working with large models in special fields where label annotation is challenging. Our code is available at https://github.com/chenjzBUAA/SATIR

arXiv:2304.07775공식 코드 · Papers with Code

Robust Cross-Modal Knowledge Distillation for Unconstrained Videos

Cross-modal distillation has been widely used to transfer knowledge across different modalities, enriching the representation of the target unimodal one. Recent studies highly relate the temporal synchronization between vision and sound to the semantic consistency for cross-modal distillation. However, such semantic consistency from the synchronization is hard to guarantee in unconstrained videos, due to the irrelevant modality noise and differentiated semantic correlation. To this end, we first propose a \textit{Modality Noise Filter} (MNF) module to erase the irrelevant noise in teacher modality with cross-modal context. After this purification, we then design a \textit{Contrastive Semantic Calibration} (CSC) module to adaptively distill useful knowledge for target modality, by referring to the differentiated sample-wise semantic correlation in a contrastive fashion. Extensive experiments show that our method could bring a performance boost compared with other distillation methods in both visual action recognition and video retrieval task. We also extend to the audio tagging task to prove the generalization of our method. The source code is available at \href{https://github.com/GeWu-Lab/cross-modal-distillation}{https://github.com/GeWu-Lab/cross-modal-distillation}.

arXiv:2304.07593공식 코드 · Papers with Code

Teacher Network Calibration Improves Cross-Quality Knowledge Distillation

We investigate cross-quality knowledge distillation (CQKD), a knowledge distillation method where knowledge from a teacher network trained with full-resolution images is transferred to a student network that takes as input low-resolution images. As image size is a deciding factor for the computational load of computer vision applications, CQKD notably reduces the requirements by only using the student network at inference time. Our experimental results show that CQKD outperforms supervised learning in large-scale image classification problems. We also highlight the importance of calibrating neural networks: we show that with higher temperature smoothing of the teacher's output distribution, the student distribution exhibits a higher entropy, which leads to both, a lower calibration error and a higher network accuracy.

arXiv:2304.07258공식 코드 · Papers with Code

Learn What Is Possible, Then Choose What Is Best: Disentangling One-To-Many Relations in Language Through Text-based Games

Language models pre-trained on large self-supervised corpora, followed by task-specific fine-tuning has become the dominant paradigm in NLP. These pre-training datasets often have a one-to-many structure--e.g. in dialogue there are many valid responses for a given context. However, only some of these responses will be desirable in our downstream task. This raises the question of how we should train the model such that it can emulate the desirable behaviours, but not the undesirable ones. Current approaches train in a one-to-one setup--only a single target response is given for a single dialogue context--leading to models only learning to predict the average response, while ignoring the full range of possible responses. Using text-based games as a testbed, our approach, PASA, uses discrete latent variables to capture the range of different behaviours represented in our larger pre-training dataset. We then use knowledge distillation to distil the posterior probability distribution into a student model. This probability distribution is far richer than learning from only the hard targets of the dataset, and thus allows the student model to benefit from the richer range of actions the teacher model has learned. Results show up to 49% empirical improvement over the previous state-of-the-art model on the Jericho Walkthroughs dataset.

arXiv:2304.05057공식 코드 · Papers with Code

SFT-KD-Recon: Learning a Student-friendly Teacher for Knowledge Distillation in Magnetic Resonance Image Reconstruction

Deep cascaded architectures for magnetic resonance imaging (MRI) acceleration have shown remarkable success in providing high-quality reconstruction. However, as the number of cascades increases, the improvements in reconstruction tend to become marginal, indicating possible excess model capacity. Knowledge distillation (KD) is an emerging technique to compress these models, in which a trained deep teacher network is used to distill knowledge to a smaller student network such that the student learns to mimic the behavior of the teacher. Most KD methods focus on effectively training the student with a pre-trained teacher unaware of the student model. We propose SFT-KD-Recon, a student-friendly teacher training approach along with the student as a prior step to KD to make the teacher aware of the structure and capacity of the student and enable aligning the representations of the teacher with the student. In SFT, the teacher is jointly trained with the unfolded branch configurations of the student blocks using three loss terms - teacher-reconstruction loss, student-reconstruction loss, and teacher-student imitation loss, followed by KD of the student. We perform extensive experiments for MRI acceleration in 4x and 5x under-sampling on the brain and cardiac datasets on five KD methods using the proposed approach as a prior step. We consider the DC-CNN architecture and setup teacher as D5C5 (141765 parameters), and student as D3C5 (49285 parameters), denoting a compression of 2.87:1. Results show that (i) our approach consistently improves the KD methods with improved reconstruction performance and image quality, and (ii) the student distilled using our approach is competitive with the teacher, with the performance gap reduced from 0.53 dB to 0.03 dB.

arXiv:2304.03980공식 코드 · Papers with Code

Continual Learning for LiDAR Semantic Segmentation: Class-Incremental and Coarse-to-Fine strategies on Sparse Data

During the last few years, continual learning (CL) strategies for image classification and segmentation have been widely investigated designing innovative solutions to tackle catastrophic forgetting, like knowledge distillation and self-inpainting. However, the application of continual learning paradigms to point clouds is still unexplored and investigation is required, especially using architectures that capture the sparsity and uneven distribution of LiDAR data. The current paper analyzes the problem of class incremental learning applied to point cloud semantic segmentation, comparing approaches and state-of-the-art architectures. To the best of our knowledge, this is the first example of class-incremental continual learning for LiDAR point cloud semantic segmentation. Different CL strategies were adapted to LiDAR point clouds and tested, tackling both classic fine-tuning scenarios and the Coarse-to-Fine learning paradigm. The framework has been evaluated through two different architectures on SemanticKITTI, obtaining results in line with state-of-the-art CL strategies and standard offline learning.