检索结果-内蒙古大学图书馆

Local Action-Guided Motion Diffusion Model for Text-to-Motion Generation

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Jin, Peng Li, Hao Cheng, Zesen Li, Kehan Yu, Runyi Liu, Chang Ji, Xiangyang Yuan, Li Chen, Jie School of Electronic and Computer Engineering Peking University Shenzhen China Peng Cheng Laboratory Shenzhen China -Preferred Program Peking University Shenzhen Graduate School Shenzhen China Department of Automation and BNRist Tsinghua University Beijing China

Text-to-motion generation requires not only grounding local actions in language but also seamlessly blending these individual actions to synthesize diverse and realistic global motions. However, existing motion generation methods primarily focus on the direct synthesis of global motions while neglecting the importance of generating and controlling local actions. In this paper, we propose the local action-guided motion diffusion model, which facilitates global motion generation by utilizing local actions as fine-grained control signals. Specifically, we provide an automated method for reference local action sampling and leverage graph attention networks to assess the guiding weight of each local action in the overall motion synthesis. During the diffusion process for synthesizing global motion, we calculate the local-action gradient to provide conditional guidance. This local-to-global paradigm reduces the complexity associated with direct global motion generation and promotes motion diversity via sampling diverse actions as conditions. Extensive experiments on two human motion datasets, i.e., HumanML3D and KIT, demonstrate the effectiveness of our method. Furthermore, our method provides flexibility in seamlessly combining various local actions and continuous guiding weight adjustment, accommodating diverse user preferences, which may hold potential significance for the community. The project page is available at https://***/GuidedMotion-project/. Copyright © 2024, The Authors. All rights reserved.

关键词： Blending

TG-VQA: Ternary Game of Video Question Answering

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Li, Hao Jin, Peng Cheng, Zesen Zhang, Songyang Chen, Kai Wang, Zhennan Liu, Chang Chen, Jie School of Electronic and Computer Engineering Peking University Shenzhen Graduate School Shenzhen China Shanghai AI Laboratory Shanghai China -Preferred Program Peking University Shenzhen Graduate School China Peng Cheng Laboratory Shenzhen China Department of Automation and BNRist Tsinghua University China

Video question answering aims at answering a question about the video content by reasoning the alignment semantics within them. However, since relying heavily on human instructions, i.e., annotations or priors, current contrastive learning-based VideoQA methods remains challenging to perform fine-grained visual-linguistic alignments. In this work, we innovatively resort to game theory, which can simulate complicated relationships among multiple players with specific interaction strategies, e.g., video, question, and answer as ternary players, to achieve fine-grained alignment for VideoQA task. Specifically, we carefully design a VideoQA-specific interaction strategy to tailor the characteristics of VideoQA, which can mathematically generate the fine-grained visual-linguistic alignment label without label-intensive efforts. Our TG-VQA outperforms existing state-of-the-art by a large margin (more than 5%) on long-term and short-term VideoQA datasets, verifying its effectiveness and generalization ability. Thanks to the guidance of game-theoretic interaction, our model impressively convergences well on limited data (104 videos), surpassing most of those pre-trained on large-scale data (107 videos). © 2023, CC BY.

关键词： Alignment

Text-Video Retrieval with Disentangled Conceptualization and Set-to-Set Alignment

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Jin, Peng Li, Hao Cheng, Zesen Huang, Jinfa Wang, Zhennan Yuan, Li Liu, Chang Chen, Jie School of Electronic and Computer Engineering Peking University Shenzhen China -Preferred Program Peking University Shenzhen Graduate School Shenzhen China Peng Cheng Laboratory Shenzhen China Department of Automation and BNRist Tsinghua University Beijing China

Text-video retrieval is a challenging cross-modal task, which aims to align visual entities with natural language descriptions. Current methods either fail to leverage the local details or are computationally expensive. What’s worse, they fail to leverage the heterogeneous concepts in data. In this paper, we propose the Disentangled Conceptualization and Set-to-set Alignment (DiCoSA) to simulate the conceptualizing and reasoning process of human beings. For disentangled conceptualization, we divide the coarse feature into multiple latent factors related to semantic concepts. For set-to-set alignment, where a set of visual concepts correspond to a set of textual concepts, we propose an adaptive pooling method to aggregate semantic concepts to address the partial matching. In particular, since we encode concepts independently in only a few dimensions, DiCoSA is superior at efficiency and granularity, ensuring fine-grained interactions using a similar computational complexity as coarse-grained alignment. Extensive experiments on five datasets, including MSRVTT, LSMDC, MSVD, ActivityNet, and DiDeMo, demonstrate that our method outperforms the existing state-of-the-art methods. Copyright © 2023, The Authors. All rights reserved.

关键词： Alignment

DiffusionRet: Generative Text-Video Retrieval with Diffusion Model

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Jin, Peng Li, Hao Cheng, Zesen Li, Kehan Ji, Xiangyang Liu, Chang Yuan, Li Chen, Jie School of Electronic and Computer Engineering Peking University Shenzhen China Peng Cheng Laboratory Shenzhen China -Preferred Program Peking University Shenzhen Graduate School Shenzhen China Department of Automation and BNRist Tsinghua University Beijing China

Existing text-video retrieval solutions are, in essence, discriminant models focused on maximizing the conditional likelihood, i.e., p(candidates|query). While straightforward, this de facto paradigm overlooks the underlying data distribution p(query), which makes it challenging to identify out-of-distribution data. To address this limitation, we creatively tackle this task from a generative viewpoint and model the correlation between the text and the video as their joint probability p(candidates, query). This is accomplished through a diffusion-based text-video retrieval framework (DiffusionRet), which models the retrieval task as a process of gradually generating joint distribution from noise. During training, DiffusionRet is optimized from both the generation and discrimination perspectives, with the generator being optimized by generation loss and the feature extractor trained with contrastive loss. In this way, DiffusionRet cleverly leverages the strengths of both generative and discriminative methods. Extensive experiments on five commonly used text-video retrieval benchmarks, including MSRVTT, LSMDC, MSVD, ActivityNet Captions, and DiDeMo, with superior performances, justify the efficacy of our method. More encouragingly, without any modification, DiffusionRet even performs well in out-domain retrieval settings. We believe this work brings fundamental insights into the related fields. Code is available at https://***/jpthu17/DiffusionRet. Copyright © 2023, The Authors. All rights reserved.

关键词： Machine learning

Expectation-maximization contrastive learning for compact video-and-language representations 22

学校读者我要写书评

暂无评论

Expectation-maximization contrastive learning for compact vi...

Proceedings of the 36th International Conference on Neural Information Processing Systems

作者： Peng Jin Jinfa Huang Fenglin Liu Xian Wu Shen Ge Guoli Song David A. Clifton Jie Chen School of Electronic and Computer Engineering Peking University China and AI for Science (AI4S)-Preferred Program Peking University Shenzhen Graduate School China Department of Engineering Science University of Oxford UK Tencent JARVIS Lab China Peng Cheng Laboratory Shenzhen China Department of Engineering Science University of Oxford UK and Oxford-Suzhou Centre for Advanced Research Suzhou China School of Electronic and Computer Engineering Peking University China and Peng Cheng Laboratory Shenzhen China and AI for Science (AI4S)-Preferred Program Peking University Shenzhen Graduate School China

ISBN: (纸本)9781713871088

Most video-and-language representation learning approaches employ contrastive learning, e.g., CLIP [53], to project the video and text features into a common latent space according to the semantic similarities of text-video pairs. However, such learned shared latent spaces are not often optimal, and the modality gap between visual and textual representation can not be fully eliminated. In this paper, we propose Expectation-Maximization Contrastive Learning (EMCL) to learn compact video-and-language representations. Specifically, we use the Expectation-Maximization algorithm to find a compact set of bases for the latent space, where the features could be concisely represented as the linear combinations of these bases. Such feature decomposition of video-and-language representations reduces the rank of the latent space, resulting in increased representing power for the semantics. Extensive experiments on three benchmark text-video retrieval datasets prove that our EMCL can learn more discriminative video-and-language representations than previous methods, and significantly outperform previous state-of-the-art methods across all metrics. More encouragingly, the proposed method can be applied to boost the performance of existing approaches either as a jointly training layer or an out-of-the-box inference module with no extra training, making it easy to be incorporated into any existing methods.

关键词：

Roadmap on Nonlocality in Photonic Materials and Metamaterials

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Monticone, Francesco Mortensen, N. Asger Fernández-Domínguez, Antonio I. Luo, Yu Tserkezis, Christos Khurgin, Jacob B. Shahbazyan, Tigran V. Chaves, André J. Peres, Nuno M.R. Wegner, Gino Busch, Kurt Hu, Huatian della Sala, Fabio Zhang, Pu Ciracì, Cristian Aizpurua, Javier Babaze, Antton Borisov, Andrei G. Chen, Xue-Wen Christensen, Thomas Yan, Wei Yang, Yi Hohenester, Ulrich Huber, Lorenz Wubs, Martijn de Liberato, Simone Gonçalves, P.A.D. de Abajo, F. Javier García Hess, Ortwin Tarasenko, Illya Cox, Joel D. Jelver, Line Dias, Eduardo J.C. Sánchez, Miguel Sánchez Margetis, Dionisios Gómez-Santos, Guillermo Stauber, Tobias Tretyakov, Sergei Simovski, Constantin Pakniyat, Samaneh Gómez-Díaz, J. Sebastián Bondarev, Igor V. Biehs, Svend-Age Boltasseva, Alexandra Shalaev, Vladimir M. Krasavin, Alexey V. Zayats, Anatoly V. Alù, Andrea Song, Jung-Hwan Brongersma, Mark L. Levy, Uriel Long, Olivia Y. Guo, Cheng Fan, Shanhui Bozhevolnyi, Sergey I. Overvig, Adam Prudêncio, Filipa R. Silveirinha, Mário G. Gangaraj, S. Ali Hassani Argyropoulos, Christos Huidobro, Paloma A. Galiffi, Emanuele Yang, Fan Pendry, John B. Miller, David A.B. School of Electrical and Computer Engineering Cornell University IthacaNY14853 United States POLIMA Center for Polariton-driven Light–Matter Interactions University of Southern Denmark Campusvej 55 Odense MDK-5230 Denmark D-IAS—Danish Institute for Advanced Study University of Southern Denmark Campusvej 55 Odense MDK-5230 Denmark Departamento de Física Teórica de la Materia Condensada Universidad Autónoma de Madrid MadridE-28049 Spain Universidad Autónoma de Madrid MadridE-28049 Spain School of Electrical and Electronic Engineering Nanyang Technological University Singapore639798 Singapore Key Laboratory of Radar Imaging and Microwave Photonics Ministry of Education College of Electronic and Information Engineering Nanjing University of Aeronautics and Astronautics Nanjing211106 China Department of Electrical and Computer Engineering Johns Hopkins University BaltimoreMD21218 United States Department of Physics Jackson State University JacksonMS39217 United States Department of Physics Aeronautics Institute of Technology SP So Jos dos Campos12228-900 Brazil Departamento de Física Universidade do Minho BragaP-4710-057 Portugal Av Mestre José Veiga Braga4715-330 Portugal Max-Born-Institut Berlin12489 Germany Humboldt-Universität zu Berlin Institut für Physik AG Theoretische Optik and Photonik Berlin12489 Germany Center for Biomolecular Nanotechnologies Istituto Italiano di Tecnologia Via Barsanti 14 Arnesano73010 Italy Via Monteroni Campus Unisalento Lecce73100 Italy School of physics Huazhong University of Science and Technology Luoyu Road 1037 Wuhan430074 China Donostia International Physics Center DIPC Donostia-San Sebastián20018 Spain Ikerbasque Basque Foundation for Science Bilbao48009 Spain Department of Electricity and Electronics University of the Basque Country Leioa48940 Spain Department of Applied Physics University of the Basque Country Donostia20018 Spain Université Paris-Saclay CNRS Institut des Scienc

Photonic technologies continue to drive the quest for new optical materials with unprecedented responses. A major frontier in this field is the exploration of nonlocal (spatially dispersive) materials, going beyond the local, wavevector-independent assumption traditionally made in optical material modeling. On one end, the growing interest in plasmonic, polaritonic and quantum materials has revealed naturally occurring nonlocalities, emphasizing the need for more accurate models to predict and design their optical responses. This has major implications also for topological, nonreciprocal, and time-varying systems based on these material platforms. Beyond natural materials, artificially structured materials—metamaterials and metasurfaces–can provide even stronger and engineered nonlocal effects, emerging from long-range interactions or multipolar effects. This is a rapidly expanding area in the field of photonic metamaterials, with open frontiers yet to be explored. In the case of metasurfaces, in particular, nonlocality engineering has become a powerful tool for designing strongly wavevector-dependent responses, enabling enhanced wavefront control, spatial compression, multifunctional devices, and wave-based computing. Furthermore, nonlocality and related concepts play a critical role in defining the ultimate limits of what is possible in optics, photonics, and wave physics. This Roadmap aims to survey the most exciting developments in nonlocal photonic materials, highlight new opportunities and open challenges, and chart new pathways that will drive this emerging field forward–toward new scientific discoveries and technological advancements. Copyright © 2025, The Authors. All rights reserved.

关键词： Nanocrystals

A study on phase noise characteristics of oscillation circuits focused signal output position

学校读者我要写书评

暂无评论

IEEJ Transactions on Electronics, Information and Systems

IEEJ Transactions on electronics, Information and Systems 2008年第8期128卷 1377-1378页

作者： Imaike, Takeshi Sakuta, Yukinori Sekine, Yoshifumi Program in Electronic Engineering Graduate School of Science and Technology Nihon University 7-24-1 Narashinodai Funabashi Chiba 274-8501 Japan Department of Electronics and Computer Science College of Science and Technology Nihon University 7-24-1 Narashinodai Funabashi Chiba 274-8501 Japan

This paper discusses phase noise characteristics of an Fr oscillator focusing on a signal output position. Oscillation circuits can be divided into an amplifier and a feedback circuit. We compared that phase noise characteristics before passing through a feedback circuit or a after. As a result, it turned out that the slope of the phase noise characteristic does not relate to the signal output position. © 2008 The Institute of Electrical Engineers of Japan.

关键词： Phase noise

Data mining technique with cluster anaysis use K-means algorithm for LQ45 index on Indonesia stock exchange 1

学校读者我要写书评

暂无评论

Data mining technique with cluster anaysis use K-means algor...

1st International Conference on Information and Communications Technology, ICOIACT 2018

作者： Condrobimo, A. Raharto Abbas, Bahtiar Saleh Trisetyarso, Agung Suparta, Wayan Kang, Chul-Ho Computer Science Department BINUS Graduate Program - Doctor of Computer Science Bina Nusantara University Jakarta11480 Indonesia Information System Department School of Information System Bina Nusantara University Jakarta11480 Indonesia Civil Engineering Department University of Technology Yogyakarta Yogyakarta Indonesia Department of Electronic and Communication Engineering Kwangwoon University Korea Republic of

ISBN: (纸本)9781538609545

This study aims to apply data mining techniques with cluster analysis on stock data registered in LQ45 in Indonesia Stock Exchange. The cluster analysis used in this method is k-means algorithm, the data in this research is taken from Indonesia Stock Exchange. The cluster analysis in this study analyzed the characteristics of data volumes and stock values, while the results in this study were presented in the form of cluster members visually. Therefore, this cluster analysis in this research can be used for quick and efficient identifier for each member of LQ45 index cluster based on share value for each cluster and its volume. The identification results can be used by beginner-level investors that begun to be interested in stock investments to help make informed decisions about stock trading on desired cluster groups. © 2018 IEEE.

关键词： Cluster analysis

A Multi-phase Deep Learning Approach for Predicting HCC Response to TACE using Complete Computed Tomography

学校读者我要写书评

暂无评论

A Multi-phase Deep Learning Approach for Predicting HCC Resp...

Data Science and engineering in Healthcare, Medicine and Biology, IEEE EMBS Special Topic Conference on

作者： Shun Cheng Chang Chien Fu Liu Marta Misztal Yi Hsien Hsieh Che Lin Graduate Institute of Communication Engineering National Taiwan University (NTU) Taiwan School of Electronic Engineering and Computer Science Queen Mary University of London UK Department of Electrical Engineering NTU Taiwan Center for Computational and Systems Biology and Center for Biotechnology NTU Taiwan Smart Medicine and Health Informatics Program NTU Taiwan

In this research study, we compare the predictive performance of two advanced deep learning-based models in order to provide a solution to TACE (Transarterial Chemoembolization) response prediction in HCC (Hepatocellular Carcinoma) patients. Using entire abdominal CT scans enabled a broader perspective available for the model, eliminating the need for segmentation during the preprocessing. Making use of both single-phase and multi-phase CT imaging, we have used DenseNet121 and have obtained an accuracy of 80% for the multi-phase *** Relevance: The ability to predict the effectiveness of TACE treatment prior to its administration makes it possible to provide a better decision-making aid for physicians and patients.

关键词：