检索结果-内蒙古大学图书馆

Learning Multi-expert Distribution Calibration for Long-tailed Video Classification

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Hu, Yufan Gao, Junyu Xu, Changsheng University of Science and Technology Beijing Beijing100083 China Peng Cheng Laboratory Shenzhen518055 China National Lab of Pattern Recognition Institute of Automation Chinese Academy of Sciences Beijing100190 China School of Artificial Intelligence University of Chinese Academy of Sciences Beijing China

Most existing state-of-the-art video classification methods assume that the training data obey a uniform distribution. However, video data in the real world typically exhibit an imbalanced long-tailed class distribution, resulting in a model bias towards head class and relatively low performance on tail class. While the current long-tailed classification methods usually focus on image classification, adapting it to video data is not a trivial extension. We propose an end-to-end multiexpert distribution calibration method to address these challenges based on two-level distribution information. The method jointly considers the distribution of samples in each class (intra-class distribution) and the overall distribution of diverse data (interclass distribution) to solve the issue of imbalanced data under long-tailed distribution. By modeling the two-level distribution information, the model can jointly consider the head classes and the tail classes and significantly transfer the knowledge from the head classes to improve the performance of the tail classes. Extensive experiments verify that our method achieves state-ofthe-art performance on the long-tailed video classification task. Copyright © 2022, The Authors. All rights reserved.

关键词： Calibration

UniFormer: Unifying Convolution and Self-attention for Visual recognition

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Li, Kunchang Wang, Yali Zhang, Junhao Gao, Peng Song, Guanglu Liu, Yu Li, Hongsheng Qiao, Yu ShenZhen Key Lab of Computer Vision and Pattern Recognition Shenzhen Institute of Advanced Technology Chinese Academy of Sciences Shenzhen518055 China University of Chinese Academy of Sciences Beijing100049 China Shanghai Artificial Intelligence Laboratory Shanghai200232 China National University of Singapore Singapore Shanghai Artificial Intelligence Laboratory China SenseTime Research China The Chinese University of Hong Kong Hong Kong

It is a challenging task to learn discriminative representation from images and videos, due to large local redundancy and complex global dependency in these visual data. Convolution neural networks (CNNs) and vision transformers (ViTs) have been two dominant frameworks in the past few years. Though CNNs can efficiently decrease local redundancy by convolution within a small neighborhood, the limited receptive field makes it hard to capture global dependency. Alternatively, ViTs can effectively capture long-range dependency via self-attention, while blind similarity comparisons among all the tokens lead to high redundancy. To resolve these problems, we propose a novel Unified transFormer (UniFormer), which can seamlessly integrate the merits of convolution and self-attention in a concise transformer format. Different from the typical transformer blocks, the relation aggregators in our UniFormer block are equipped with local and global token affinity respectively in shallow and deep layers, allowing tackling both redundancy and dependency for efficient and effective representation learning. Finally, we flexibly stack our blocks into a new powerful backbone, and adopt it for various vision tasks from image to video domain, from classification to dense prediction. Without any extra training data, our UniFormer achieves 86.3 top-1 accuracy on ImageNet-1K classification task. With only ImageNet-1K pre-training, it can simply achieve state-of-the-art performance in a broad range of downstream tasks. It obtains 82.9/84.8 top-1 accuracy on Kinetics-400/600, 60.9/71.2 top-1 accuracy on Something-Something V1/V2 video classification tasks, 53.8 box AP and 46.4 mask AP on COCO object detection task, 50.8 mIoU on ADE20K semantic segmentation task, and 77.4 AP on COCO pose estimation task. Moreover, we build an efficient UniFormer with a concise hourglass design of token shrinking and recovering, which achieves 2-4× higher throughput than the recent lightweight models. Code is av

关键词： Convolution

TLRM: Task-level Relation Module for GNN-based Few-Shot Learning

学校读者我要写书评

暂无评论

TLRM: Task-level Relation Module for GNN-based Few-Shot Lear...

IEEE Visual Communications and Image Processing (VCIP)

作者： Yurong Guo Zhanyu Ma Xiaoxu Li Yuan Dong Pattern Recognition and Intelligent System Lab. Beijing University of Posts and Telecommunications Beijing China Beijing Academy of Artificial Intelligence Beijing China Lanzhou University of Technology Lanzhou China

ISBN: (纸本)9781728173221

Recently, graph neural networks (GNNs) have shown powerful ability to handle few-shot classification problem, which aims at classifying unseen samples when trained with limited labeled samples per class. GNN-based few-shot learning architectures mostly replace traditional metric with a learnable GNN. In the GNN, the nodes are set as the samples' embedding, and the relationship between two connected nodes can be obtained by a network, the input of which is the difference of their embedding features. We consider this method of measuring relation of samples only models the sample-to-sample relation, while neglects the specificity of different tasks. That is, this method of measuring relation does not take the task-level information into account. To this end, we propose a new relation measure method, namely the task-level relation module (TLRM), to explicitly model the task-level relation of one sample to all the others. The proposed module captures the relation representations between nodes by considering the sample-to-task instead of sample-to-sample embedding features. We conducted extensive experiments on four benchmark datasets: mini-ImageNet, tiered-ImageNet, CUB-200-2011, and CIFAR-FS. Experimental results demonstrate that the proposed module is effective for GNN-based few-shot learning.

关键词： Measurement Visual communication Image processing Benchmark testing Graph neural networks Task analysis

Fetal Re-Identification in Multiple Pregnancy Ultrasound Images Using Deep Learning

学校读者我要写书评

暂无评论

Fetal Re-Identification in Multiple Pregnancy Ultrasound Ima...

Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC)

作者： Elisabeth Gabler Michael Nissen Thomas R. Altstidl Adriana Titzmann Kai Packhäuser Andreas Maier Peter A. Fasching Bjoern M. Eskofier Heike Leutheuser Department Artificial Intelligence in Biomedical Engineering Machine Learning and Data Analytics (MaD) Lab Friedrich-Alexander-Universität Erlangen-Nurnberg (FAU) Erlangen Germany Department of Gynecology and Obstetrics Erlangen University Hospital Friedrich-Alexander-Universität Erlangen-Nurnberg (FAU) Erlangen Germany Department of Computer Science Pattern Recognition Lab Friedrich-Alexander-Universität Erlangen-Nurnberg (FAU) Erlangen Germany

Ultrasound examinations during pregnancy can detect abnormal fetal development, which is a leading cause of perinatal mortality. In multiple pregnancies, the position of the fetuses may change between examinations. The individual fetus cannot be clearly identified. Fetal re-identification may improve diagnostic capabilities by tracing individual fetal changes. This work evaluates the feasibility of fetal re-identification on FETAL_PLANES_DB, a publicly available dataset of singleton pregnancy ultrasound images. Five dataset subsets with 6,491 images from 1,088 pregnant women and two re-identification frameworks (Torchreid, FastReID) are evaluated. FastReID achieves a mean average precision of 68.77% (68.42%) and mean precision at rank 10 score of 89.60% (95.55%) when trained on images showing the fetal brain (abdomen). Visualization with gradient-weighted class activation mapping shows that the classifiers appear to rely on anatomical features. We conclude that fetal re-identification in ultrasound images may be feasible. However, more work on additional datasets, including images from multiple pregnancies and several subsequent examinations, is required to ensure and investigate performance stability and *** relevance— To date, fetuses in multiple pregnancies cannot be distinguished between ultrasound examinations. This work provides the first evidence for feasibility of fetal re-identification in pregnancy ultrasound images. This may improve diagnostic capabilities in clinical practice in the future, such as longitudinal analysis of fetal changes or abnormalities.

关键词：

Continual relation learning via episodic memory activation and reconsolidation 58

学校读者我要写书评

暂无评论

Continual relation learning via episodic memory activation a...

58th Annual Meeting of the Association for Computational Linguistics, ACL 2020

作者： Han, Xu Dai, Yi Gao, Tianyu Lin, Yankai Liu, Zhiyuan Li, Peng Sun, Maosong Zhou, Jie State Key Lab on Intelligent Technology and Systems Institute for Artificial Intelligence Department of Computer Science and Technology Tsinghua University Beijing China Pattern Recognition Center WeChat AI Tencent Inc. China

ISBN: (纸本)9781952148255

Continual relation learning aims to continually train a model on new data to learn incessantly emerging novel relations while avoiding catastrophically forgetting old relations. Some pioneering work has proved that storing a handful of historical relation examples in episodic memory and replaying them in subsequent training is an effective solution for such a challenging problem. However, these memory-based methods usually suffer from overfitting the few memorized examples of old relations, which may gradually cause inevitable confusion among existing relations. Inspired by the mechanism in human long-term memory formation, we introduce episodic memory activation and reconsolidation (EMAR) to continual relation learning. Every time neural models are activated to learn both new and memorized data, EMAR utilizes relation prototypes for memory reconsolidation exercise to keep a stable understanding of old relations. The experimental results show that EMAR could get rid of catastrophically forgetting old relations and outperform the state-of-the-art continual learning models. The code and datasets are released on https://***/thunlp/ContinualRE. © 2020 Association for Computational Linguistics

关键词： Chemical activation

Selecting Stickers in open-Domain Dialogue through Multitask Learning

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Zhang, Zhexin Zhu, Yeshuang Fei, Zhengcong Zhang, Jinchao Zhou, Jie The CoAI Group DCST China Institute for Artificial Intelligence China State Key Lab of Intelligent Technology and Systems China Beijing National Research Center for Information Science and Technology China Tsinghua University Beijing100084 China Pattern Recognition Center WeChat AI Tencent Inc China

With the increasing popularity of online chatting, stickers are becoming important in our online communication. Selecting appropriate stickers in open-domain dialogue requires a comprehensive understanding of both dialogues and stickers, as well as the relationship between the two types of modalities. To tackle these challenges, we propose a multitask learning method comprised of three auxiliary tasks to enhance the understanding of dialogue history, emotion and semantic meaning of stickers. Extensive experiments conducted on a recent challenging dataset show that our model can better combine the multimodal information and achieve significantly higher accuracy over strong baselines. Ablation study further verifies the effectiveness of each auxiliary task. Our code is available at https://***/nonstopfor/Sticker-Selection. Copyright © 2022, The Authors. All rights reserved.

关键词： Semantics

CTRLEval: An Unsupervised Reference-Free Metric for Evaluating Controlled Text Generation

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Ke, Pei Zhou, Hao Lin, Yankai Li, Peng Zhou, Jie Zhu, Xiaoyan Huang, Minlie The CoAI group DCST Institute for Artificial Intelligence State Key Lab of Intelligent Technology and Systems Beijing National Research Center for Information Science and Technology Tsinghua University Beijing100084 China Pattern Recognition Center WeChat AI Tencent Inc. China Tsinghua University China

Existing reference-free metrics have obvious limitations for evaluating controlled text generation models. Unsupervised metrics can only provide a task-agnostic evaluation result which correlates weakly with human judgments, whereas supervised ones may over-fit task-specific data with poor generalization ability to other datasets. In this paper, we propose an unsupervised reference-free metric called CTRLEval, which evaluates controlled text generation from different aspects by formulating each aspect into multiple text infilling tasks. On top of these tasks, the metric assembles the generation probabilities from a pre-trained language model without any model training. Experimental results show that our metric has higher correlations with human judgments than other baselines, while obtaining better generalization of evaluating generated texts from different models and with different qualities Copyright © 2022, The Authors. All rights reserved.

关键词：

Active Universal Domain Adaptation

学校读者我要写书评

暂无评论

Active Universal Domain Adaptation

International Conference on Computer Vision (ICCV)

作者： Xinhong Ma Junyu Gao Changsheng Xu National Lab of Pattern Recognition (NLPR) Institute of Automation Chinese Academy of Sciences (CASIA) School of Artificial Intelligence University of Chinese Academy of Sciences (UCAS) Peng Cheng Laboratory Shenzhen China

ISBN: (纸本)9781665428132

Most unsupervised domain adaptation methods rely on rich prior knowledge about the source-target label set relationship, and they cannot recognize categories beyond the source classes, which limits their applicability in practical scenarios. This paper proposes a new paradigm for unsupervised domain adaptation, termed as Active Universal Domain Adaptation (AUDA), which removes all label set assumptions and aims for not only recognizing target samples from source classes but also inferring those from target-private classes by using active learning to annotate a small budget of target data. For AUDA, it is challenging to jointly adapt the model to the target domain and select informative target samples for annotations under a large domain gap and significant semantic shift. To address the problems, we propose an Active Universal Adaptation Network (AUAN). Specifically, we first introduce Adversarial and Diverse Curriculum Learning (ADCL), which progressively aligns source and target domains to classify whether target samples are from source classes. Then, we propose a Clustering Non-transferable Gradient Embedding (CNTGE) strategy, which utilizes the clues of transferability, diversity, and uncertainty to annotate target informative sample, making it possible to infer labels for target samples of target-private classes. Finally, we propose to jointly train ADCL and CNTGE with target supervision to promote domain adaptation and target-private class recognition. Extensive experiments demonstrate that the proposed AUDA model equipped with ADCL and CNTGE achieves significant results on four popular benchmarks.

关键词： Adaptation models Computer vision Uncertainty Target recognition Annotations Computational modeling Semantics

Does proprietary software still offer protection of intellectual property in the age of machine learning? A case study using dual energy CT data

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Maier, Andreas Yang, Seung Hee Maleki, Farhad Muthukrishnan, Nikesh Forghani, Reza Pattern Recognition Lab FAU Erlangen-Nürnberg Germany Department Artificial Intelligence in Medical Engineering FAU Erlangen-Nürnberg Germany McGill University Hospital McGill University Canada

In the domain of medical image processing, medical device manufacturers protect their intellectual property in many cases by shipping only compiled software, i.e. binary code which can be executed but is difficult to be understood by a potential attacker. In this paper, we investigate how well this procedure is able to protect image processing algorithms. In particular, we investigate whether the computation of mono-energetic images and iodine maps from dual energy CT data can be reverse-engineered by machine learning methods. Our results indicate that both can be approximated using only one single slice image as training data at a very high accuracy with structural similarity greater than 0.98 in all investigated cases. © 2021, CC BY.

关键词： Intellectual property