检索结果-内蒙古大学图书馆

DPA-2:a large atomic model as a multitask learner

npj Computational Materials 2024年第1期10卷 185-199页

作者： Duo Zhang Xinzijian Liu Xiangyu Zhang Chengqian Zhang Chun Cai Hangrui Bi Yiming Du Xuejian Qin Anyang Peng Jiameng Huang Bowen Li Yifan Shan Jinzhe Zeng Yuzhi Zhang Siyuan Liu Yifan Li Junhan Chang Xinyan Wang Shuo Zhou Jianchuan Liu Xiaoshan Luo Zhenyu Wang Wanrun Jiang Jing Wu Yudi Yang Jiyuan Yang Manyi Yang Fu-Qiang Gong Linshuang Zhang Mengchao Shi Fu-Zhi Dai Darrin M.York Shi Liu Tong Zhu Zhicheng Zhong Jian Lv Jun Cheng Weile Jia Mohan Chen Guolin Ke Weinan E Linfeng Zhang Han Wang AI for Science Institute BeijingP.R.China DP Technology BeijingP.R.China Academy for Advanced Interdisciplinary Studies Peking UniversityBeijingP.R.China State Key Lab of Processors Institute of Computing TechnologyChinese Academy of SciencesBeijingP.R.China University of Chinese Academy of Sciences BeijingP.R.China HEDPS CAPTCollege of EngineeringPeking UniversityBeijingP.R.China Ningbo Institute of Materials Technology and Engineering Chinese Academy of SciencesNingboP.R.China CAS Key Laboratory of Magnetic Materials and Devices and Zhejiang Province Key Laboratory of Magnetic Materials and Application Technology Chinese Academy of SciencesNingboP.R.China School of Electronics Engineering and Computer Science Peking UniversityBeijingP.R.China Shanghai Engineering Research Center of Molecular Therapeutics&New Drug Development School of Chemistry and Molecular EngineeringEast China Normal UniversityShanghaiP.R.China Laboratory for Biomolecular Simulation Research Institute for Quantitative Biomedicine and Department of Chemistry and Chemical BiologyRutgers UniversityPiscatawayNJUSA Department of Chemistry Princeton UniversityPrincetonNJUSA College of Chemistry and Molecular Engineering Peking UniversityBeijingP.R.China Yuanpei College Peking UniversityBeijingP.R.China School of Electrical Engineering and Electronic Information Xihua UniversityChengduP.R.China State Key Laboratory of Superhard Materials College of PhysicsJilin UniversityChangchunP.R.China Key Laboratory of Material Simulation Methods&Software of Ministry of Education College of PhysicsJilin UniversityChangchunP.R.China International Center of Future Science Jilin UniversityChangchunP.R.China Key Laboratory for Quantum Materialsof Zhejiang Province Department of PhysicsSchool of ScienceWestlake UniversityHangzhouP.R.China Atomistic Simulations Italian Institute of TechnologyGenovaItaly State Key Laboratory of Physical Chemistry of Solid Surface iChEMCollege of Chemistry and Chemical EngineeringXiame

The rapid advancements in artificial intelligence(AI)are catalyzing transformative changes in atomic modeling,simulation,and ***-driven potential energy models havedemonstrated the capability to conduct large-scale,long-duration simulations with the accuracy of ab initio electronic structure ***,the model generation process remains a bottleneck for large-scale *** propose a shift towards a model-centric ecosystem,wherein a large atomic model(LAM),pretrained across multiple disciplines,can be efficiently fine-tuned and distilled for various downstream tasks,thereby establishing a new framework for molecular *** this study,we introduce the DPA-2 architecture as a prototype for ***-trained on a diverse array of chemical and materials systemsusing a multi-task approach,DPA-2demonstrates superior generalization capabilities across multiple downstream tasks compared to the traditional single-task pre-training and fine-tuning *** approach sets the stage for the development and broad application of LAMs in molecular and materials simulation research.

关键词： DPA establishing thereby

来源：评论

学校读者我要写书评

暂无评论

You Only Align Once: Bidirectional Interaction for Spatial-Temporal Video Super-Resolution

arXiv

引用

arXiv 2022年

作者： Hu, Mengshun Jiang, Kui Nie, Zhixiang Wang, Zheng National Engineering Research Center for Multimedia Software Hubei Key Laboratory of Multimedia and Network Communication Engineering School of Computer Science Wuhan University China

Spatial-Temporal Video Super-Resolution (ST-VSR) technology generates high-quality videos with higher resolution and higher frame rates. Existing advanced methods accomplish ST-VSR tasks through the association of Spatial and Temporal video super-resolution (S-VSR and T-VSR). These methods require two alignments and fusions in S-VSR and T-VSR, which is obviously redundant and fails to sufficiently explore the information flow of consecutive spatial LR frames. Although bidirectional learning (future-to-past and past-to-future) was introduced to cover all input frames, the direct fusion of final predictions fails to sufficiently exploit intrinsic correlations of bidirectional motion learning and spatial information from all frames. We propose an effective yet efficient recurrent network with bidirectional interaction for ST-VSR, where only one alignment and fusion is needed. Specifically, it first performs backward inference from future to past, and then follows forward inference to super-resolve intermediate frames. The backward and forward inferences are assigned to learn structures and details to simplify the learning task with joint optimizations. Furthermore, a Hybrid Fusion Module (HFM) is designed to aggregate and distill information to refine spatial information and reconstruct high-quality video frames. Extensive experiments on two public datasets demonstrate that our method outperforms state-of-the-art methods in efficiency, and reduces calculation cost by about 22%. Copyright © 2022, The Authors. All rights reserved.

关键词： Optical resolving power

来源：评论

学校读者我要写书评

暂无评论

On the Effectiveness of Distillation in Mitigating Backdoors in Pre-trained Encoder

arXiv

引用

arXiv 2024年

作者： Han, Tingxu Huang, Shenghan Ding, Ziqi Sun, Weisong Feng, Yebo Fang, Chunrong Li, Jun Qian, Hanwei Wu, Cong Zhang, Quanjun Liu, Yang Chen, Zhenyu The State Key Laboratory for Novel Software Technology Nanjing University Jiangsu210093 China The School of Computer Science University of New South Wales NSW2052 Australia The School of Computer Science and Engineering Nanyang Technological University Nanyang639798 Singapore The Department of Computer Science The Director of the Network & Security Research Laboratory The University of Oregon United States

With the development of deep learning, self-supervised learning (SSL) has become particularly popular. SSL pre-trains an encoder (e.g., an image encoder) to learn generic features from a large amount of unlabeled data, releases it on a third-party platform, and allows any downstream task (e.g., an image classification task) to further fine-tune the encoder with task-specific data to derive the model for the task. However, despite the tremendous performance improvements it brings, SSL faces serious security risks. An adversary can poison pre-training datasets used to pre-train an encoder or publish a poisoned encoder directly to third-party platforms, which will also poison the model for any downstream task that uses these encoders. In this paper, we study a defense against poisoned encoders in SSL called distillation, which is a defense used in supervised learning originally. Distillation aims to distill knowledge from a given model (a.k.a the teacher net) and transfer it to another (a.k.a the student net). Now, we use it to distill benign knowledge from poisoned pre-trained encoders and transfer it to a new encoder, resulting in a clean pre-trained encoder. In particular, we conduct an empirical study on the effectiveness and performance of distillation against poisoned encoders. Using two state-of-the-art backdoor attacks against pre-trained image encoders and four commonly used image classification datasets, our experimental results show that distillation can reduce attack success rate from 80.87% to 27.51% while suffering a 6.35% loss in accuracy. Moreover, we investigate the impact of three core components of distillation on performance: Teacher net, student net, and distillation loss. By comparing 4 different teacher nets, 3 student nets, and 6 distillation losses, we find that fine-tuned teacher nets, warm-up-training-based student nets, and attention-based distillation loss perform best, respectively. Copyright © 2024, The Authors. All rights reserved.

关键词： Distillation

来源：评论

学校读者我要写书评

暂无评论

Fast target-aware learning for few-shot video object segmentation

引用

Science China(Information Sciences) 2022年第8期65卷 71-86页

作者： Yadang CHEN Chuanyan HAO Zhi-Xin YANG Enhua WU Engineering Research Center of Digital Forensics Ministry of EducationSchool of Computer and SoftwareNanjing University of Information Science and Technology School of Education Science and Technology Nanjing University of Posts and Telecommunications State Key Laboratory of Internet of Things for Smart City Department of Electromechanical EngineeringUniversity of Macau State Key Laboratory of Computer Science Institute of SoftwareUniversity of Chinese Academy of Sciences Faculty of Science and Technology University of Macau

Few-shot video object segmentation(FSVOS) aims to segment a specific object throughout a video sequence when only the first-frame annotation is given. In this study, we develop a fast target-aware learning approach for FSVOS, where the proposed approach adapts to new video sequences from its firstframe annotation through a lightweight procedure. The proposed network comprises two models. First, the meta knowledge model learns the general semantic features for the input video image and up-samples the coarse predicted mask to the original image size. Second, the target model adapts quickly from the limited support set. Concretely, during the online inference for testing the video, we first employ fast optimization techniques to train a powerful target model by minimizing the segmentation error in the first frame and then use it to predict the subsequent frames. During the offline training, we use a bilevel-optimization strategy to mimic the full testing procedure to train the meta knowledge model across multiple video *** proposed method is trained only on an individual public video object segmentation(VOS) benchmark without additional training sets and compared favorably with state-of-the-art methods on DAVIS-2017, with a J &F overall score of 71.6%, and on YouT ubeVOS-2018, with a J &F overall score of 75.4%. Meanwhile,a high inference speed of approximately 0.13 s per frame is maintained.

关键词： video object segmentation few-shot target-aware meta knowledge bilevel-optimization

来源：评论

学校读者我要写书评

暂无评论

Climate Downscaling Using Neural Operator: Spatiotemporal Multimodal Fusion Operator with state-Query Coupled Kernel

Climate Downscaling Using Neural Operator: Spatiotemporal Mu...

引用

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Haodi Zhang Yichi Wang Yifan Jian Jiahui Jiang Zhaohai Bai Lin Ma College of Computer Science and Software Engineering Shenzhen University Center for Agricultural Resources Research Institute of Genetic and Developmental Biology Chinese Academy of Sciences State Key Laboratory of Pollution Control and Resource Reuse School of the Environment Nanjing University

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Climate downscaling is crucial for detailed small- scale analysis and for acquiring climate data in regions without weather stations. Operator learning has proven potential for this task. However, several challenges remain in operator learning, such as multimodal fusion, spatiotemporal fusion and input state and query adaptation. To address these challenges, we propose a Spatiotemporal Multimodal Fusion Operator with a state- Query Coupled Kernel (SMCK). This framework includes a latent space fusion encoder that encodes climate variables using position-wise multihead attention for multimodal fusion and integrates historical information to generate robust and precise representation. Additionally, we introduce a state-query coupled kernel that combines radial basis functions and discrete fourier encoding to enhance query location representation, while also adapting to the state to obtain the coupled kernel. Extensive experiments demonstrate that our method achieves state-of-the-art performance and provides strong support for climate downscaling and the planning of climate-related strategies.

关键词： Couplings Signal processing Encoding Robustness Spatiotemporal phenomena Planning Kernel Speech processing Forecasting Meteorology

来源：评论

学校读者我要写书评

暂无评论

FedCompetitors: Harmonious Collaboration in Federated Learning with Competing Participants

arXiv

引用

arXiv 2023年

作者： Tan, Shanli Cheng, Hao Wu, Xiaohu Yu, Han He, Tiantian Ong, Yew-Soon Wang, Chongjun Tao, Xiaofeng National Engineering Research Center of Mobile Network Technologies Beijing University of Posts and Telecommunications China State Key Laboratory for Novel Software Technology Nanjing University China School of Computer Science and Engineering Nanyang Technological University Singapore Agency for Science Technology and Research Singapore

Federated learning (FL) provides a privacy-preserving approach for collaborative training of machine learning models. Given the potential data heterogeneity, it is crucial to select appropriate collaborators for each FL participant (FL-PT) based on data complementarity. Recent studies have addressed this challenge. Similarly, it is imperative to consider the inter-individual relationships among FL-PTs where some FL-PTs engage in competition. Although FL literature has acknowledged the significance of this scenario, practical methods for establishing FL ecosystems remain largely unexplored. In this paper, we extend a principle from the balance theory, namely "the friend of my enemy is my enemy", to ensure the absence of conflicting interests within an FL ecosystem. The extended principle and the resulting problem are formulated via graph theory and integer linear programming. A polynomial-time algorithm is proposed to determine the collaborators of each FL-PT. The solution guarantees high scalability, allowing even competing FL-PTs to smoothly join the ecosystem without conflict of interest. The proposed framework jointly considers competition and data heterogeneity. Extensive experiments on real-world and synthetic data demonstrate its efficacy compared to five alternative approaches, and its ability to establish efficient collaboration networks among FL-PTs. Copyright © 2023, The Authors. All rights reserved.

关键词： Ecosystems

来源：评论

学校读者我要写书评

暂无评论

EFFICIENT MULTI-STEP AUDIO OBJECT CODING WITH LIMITED RESIDUAL INFORMATION

EFFICIENT MULTI-STEP AUDIO OBJECT CODING WITH LIMITED RESIDU...

引用

2021 IEEE International Conference on Multimedia and Expo, ICME 2021

作者： Hu, Chenhao Hu, Ruimin Wang, Xiaochen Wu, Yulin Liu, Wenke National Engineering Research Center for Multimedia Software School of Computer Science Wuhan University China Research Institute of Wuhan University Shenzhen China Hubei Key Laboratory of Multimedia and Network Communication Engineering Wuhan University China

ISBN: (纸本)9781665438643

Spatial audio object coding (SAOC) is an effective method to transmit multiple audio objects. Audio systems can provide personalized services under this framework. However, this method causes frequency aliasing distortion, which severely impacts the listening experience. The multi-step SAOC (MS-SAOC) scheme was proposed to enhance the sound quality of each audio object by using residual information. Compared with SAOC, the bit-rate increases three times due to the residual data of multiple objects. In this paper, an efficient multi-step residual coding method is proposed to reduce the residual bit-rate of MS-SAOC. A two-level filter is designed to remove redundant residual information, and the limited residual information can efficiently compensate for frequency aliasing distortion. From experiment results, the residual bit-rate is half of MS-SAOC, and the sound quality is maintained at the Good-Excellent level. © 2021 IEEE

关键词： Audio systems

来源：评论

学校读者我要写书评

暂无评论

LOW BITRATES AUDIO OBJECT CODING USING CONVOLUTIONAL AUTO-ENCODER AND DENSENET MIXTURE MODEL

LOW BITRATES AUDIO OBJECT CODING USING CONVOLUTIONAL AUTO-EN...

引用

2021 IEEE International Conference on Multimedia and Expo, ICME 2021

作者： Wu, Yulin Hu, Ruimin Hu, Chenhao Ke, Shanfa Li, Gang Wang, Xiaochen National Engineering Research Center for Multimedia Software School of Computer Science Wuhan University China Hubei Key Laboratory of Multimedia and Network Communication Engineering Wuhan University China Research Institute of Wuhan University Shenzhen China

ISBN: (纸本)9781665438643

The efficient transmission of the audio objects can be achieved by spatial audio object coding (SAOC) method that conveys a mono downmix signal together with side information parameters that enable object reconstruction in the decoder. To allow the transmission of audio objects at low bitrates, we present a new audio coding method with convolutional auto-encoder (CAE) and dense convolutional network (DenseNet) mixture model, optimizing the compression of side information parameters of audio objects. It has two main advantages: 1) Different from the linear transform methods, CAE can dig the nonlinear relationship of side information parameters and can effectively reduce the dimension of side information parameters;2) DenseNet is adding in the decoder to make full use of low dimensional features of side information parameters, which improves the audio quality at low bitrate. Experiments show that our method outperforms baseline methods permitting bitrates as low as 1 kbps per object. © 2021 IEEE

关键词： Mixtures

来源：评论

学校读者我要写书评

暂无评论

SPATIAL AUDIO OBJECT CODING BASED ON TIME-FREQUENCY SHIFTING AND SCHEDULING

SPATIAL AUDIO OBJECT CODING BASED ON TIME-FREQUENCY SHIFTING...

引用

2021 IEEE International Conference on Multimedia and Expo, ICME 2021

作者： Hu, Chenhao Hu, Ruimin Wang, Xiaochen Wu, Yulin National Engineering Research Center for Multimedia Software School of Computer Science Wuhan University China Research Institute of Wuhan University Shenzhen China Hubei Key Laboratory of Multimedia and Network Communication Engineering Wuhan University China

ISBN: (纸本)9781665438643

Spatial audio object coding (SAOC) is an effective method to transmit multiple audio objects. It divides the full frequency band into 28 subbands and extracts spatial parameters for decoding. In this way, objects can be encoded into a downmix signal with a few parameters. However, using the same parameters in one subband will cause frequency aliasing distortion, which severely impacts the listening experience. Existing studies to enhance SAOC cannot eliminate the aliasing distortion of all objects effectively. This paper describes a new structure to balance the bit-rate and decode quality based on time-frequency (TF) shifting and scheduling. In this structure, a TF shifting strategy (contains global shifting and local shifting) is proposed to reduce frequency aliasing distortion. Furthermore, a scheduling strategy is used to decide which part should be shifted according to the degree of aliasing. From the experiment results, the performance of the proposed method is better than SAOC and other enhanced methods. © 2021 IEEE computer Society. All rights reserved.

关键词： Decoding

来源：评论

学校读者我要写书评

暂无评论

Illuminate Low-Light Image via Coarse-to-fine Multi-level network 27th

Illuminate Low-Light Image via Coarse-to-fine Multi-level Ne...

引用

27th International Conference on MultiMedia Modeling, MMM 2021

作者： Qiu, Yansheng Chen, Jun Wang, Xiao Jang, Kui National Engineering Research Center for Multimedia Software School of Computer Science Wuhan University Wuhan430072 China Hubei Key Laboratory of Multimedia and Network Communication Engineering Wuhan University Wuhan430072 China

ISBN: (纸本)9783030678319

Images under low light or against light are of low readability and visibility which in turn cause performance degradation of many computer vision tasks. As the crucial research objects for low-brightness image enhancement, reflectance as an invariant has color information, and illuminance as a variable has brightness information. Existing technologies prefer to represent the reflectance and illumination in the same network depth directly, which ignores the internal discrepancy and correlation of these two components, further causing the details vanished and artifacts noises. Consequently, we propose coarse-to-fine multi-level decouple and fusion network (CMDFN) to decompose the restoration tasks into two sub-tasks, and use decouple network (CMDN) to learn the joint features of reflectance and illumination at different network depths. To refine the joint representation and revise color information, fusion network (CMFN) is applied to promote the illuminations map by extracting the complementary and redundant information from the reflectance map. Finally, we integrate the refined illumination and reflectance to reproduce the predicted image. Especially, in CMFN, we introduce a process based on discriminate attention mechanism to balance the noise. Experimental results have demonstrated that our proposed network is superior other state of the art methods, both visually and quantitatively. © 2021, Springer Nature Switzerland AG.

关键词： Reflection

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：