检索结果-内蒙古大学图书馆

IEEE International Conference on Bioinformatics and Biomedicine (BIBM)

作者： Mingxin Liu Yunzan Liu Hui Cui Chunquan Li Jiquan Ma Department of Computer Science and Technology Heilongjiang University Harbin China Department of Computer Science and Information Technology La Trobe University Melbourne Australia The First Affiliated Hospital Institute of Cardiovascular Disease Hengyang Medical School University of South China Hengyang China The First Affiliated Hospital Cardiovascular Lab of Big Data and Imaging Artificial Intelligence Hengyang Medical School University of South China Hengyang China Department of Biochemistry and Molecular Biology School of Basic Medical Sciences Hengyang Medical School University of South China Hengyang China

The rapidly emerging field of deep learning-based computational pathology has shown promising results in utilizing whole slide images (WSIs) to objectively prognosticate cancer patients. However, most prognostic methods are currently limited to either histopathology or genomics alone, which inevitably reduces their potential to accurately predict patient prognosis. Whereas integrating WSIs and genomic features presents three main challenges: (1) the enormous heterogeneity of gigapixel WSIs which can reach sizes as large as 150,000×150,000 pixels; (2) the absence of a spatially corresponding relationship between histopathology images and genomic molecular data; and (3) the existing early, late, and intermediate multimodal feature fusion strategies struggle to capture the explicit interactions between WSIs and genomics. To ameliorate these issues, we propose the Mutual-Guided Cross-Modality Transformer (MGCT), a weakly-supervised, attention-based multimodal learning framework that can combine histology features and genomic features to model the genotype-phenotype interactions within the tumor microenvironment. To validate the effectiveness of MGCT, we conduct experiments using nearly 3,600 gigapixel WSIs across five different cancer types sourced from The Cancer Genome Atlas (TCGA). Extensive experimental results consistently emphasize that MGCT outperforms the state-of-the-art (SOTA) methods.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Investigating device-specific visual feedback for cross-device transfer in table-centric multisurface environments

Investigating device-specific visual feedback for cross-devi...

引用

作者： Scott, Stacey D. Besacier, Guillaume Goyal, Nippun Cento, Frank School of Computer Science University of Guelph GuelphONN1G 2W1 Canada Department of Systems Design Engineering University of Waterloo WaterlooONN2L 3G1 Canada Department of Computer Science Paris University 8 2 rue de la Liberté Saint-Denis93526 France

Table-centric multisurface environments (T-MSEs) that combine small multitouch surfaces (eg, smartphones and tablets) with large interactive tabletops provide people with both personal and shared workspaces to support various independent and collective tasks during group activities. This paper reports on the third in a series of studies exploring how existing interaction methods for cross-device transfer, such as the Pick-and-Drop (P&D) method, can be adapted to table-centric multisurface environment settings. The study examined the use of device-specific visual feedback to improve users' awareness of transferred content during P&D transfer. The tabletop feedback utilized the existing Surface Ghosts P&D feedback approach (ie, "ghosted" versions of transferred content were displayed in real time under the user's hand). The tablet feedback consisted of a static "Tablet Bridge" feedback showing miniature versions of transferred content along the top edge of the tablet interface. The study found that providing both types of feedback significantly improved users' transfer awareness over providing Surface Ghosts feedback alone. It also revealed that the Tablet Bridge feedback helped compensate for technical and usability issues associated with the Surface Ghosts feedback design. Lessons learned from our combined series of cross-device transfer studies are reflected upon, and relevant design implications are discussed. © 2017 John Wiley & Sons, Ltd.

关键词： Bridges

来源：评论

学校读者我要写书评

暂无评论

MGCT: Mutual-Guided Cross-Modality Transformer for Survival Outcome Prediction using Integrative Histopathology-Genomic Features

arXiv

引用

arXiv 2023年

作者： Liu, Mingxin Liu, Yunzan Cui, Hui Li, Chunquan Ma, Jiquan Department of Computer Science and Technology Heilongjiang University Harbin China Department of Computer Science and Information Technology La Trobe University Melbourne Australia The First Affiliated Hospital Institute of Cardiovascular Disease Hengyang Medical School University of South China Hengyang China The First Affiliated Hospital Cardiovascular Lab of Big Data and Imaging Artificial Intelligence Hengyang Medical School University of South China Hengyang China Department of Biochemistry and Molecular Biology School of Basic Medical Sciences Hengyang Medical School University of South China Hengyang China

The rapidly emerging field of deep learning-based computational pathology has shown promising results in utilizing whole slide images (WSIs) to objectively prognosticate cancer patients. However, most prognostic methods are currently limited to either histopathology or genomics alone, which inevitably reduces their potential to accurately predict patient prognosis. Whereas integrating WSIs and genomic features presents three main challenges: (1) the enormous heterogeneity of gigapixel WSIs which can reach sizes as large as 150,000×150,000 pixels;(2) the absence of a spatially corresponding relationship between histopathology images and genomic molecular data;and (3) the existing early, late, and intermediate multimodal feature fusion strategies struggle to capture the explicit interactions between WSIs and genomics. To ameliorate these issues, we propose the Mutual-Guided Cross-Modality Transformer (MGCT), a weakly-supervised, attention-based multimodal learning framework that can combine histology features and genomic features to model the genotype-phenotype interactions within the tumor microenvironment. To validate the effectiveness of MGCT, we conduct experiments using nearly 3,600 gigapixel WSIs across five different cancer types sourced from The Cancer Genome Atlas (TCGA). Extensive experimental results consistently emphasize that MGCT outperforms the state-of-the-art (SOTA) methods. © 2023, CC BY.

关键词： Genome

来源：评论

学校读者我要写书评

暂无评论

Multivariate Time Series Forecasting with Dynamic Graph Neural ODEs

arXiv

引用

arXiv 2022年

作者： Jin, Ming Zheng, Yu Li, Yuan-Fang Chen, Siheng Yang, Bin Pan, Shirui The Department of Data Science and AI Faculty of IT Monash University Clayton Australia The Department of Computer Science and Information Technology La Trobe University Melbourne Australia Shanghai Jiao Tong University Shanghai China The School of Data Science and Engineering East China Normal University China The School of Information and Communication Technology Griffith University SouthportQLD4222 Australia

Multivariate time series forecasting has long received significant attention in real-world applications, such as energy consumption and traffic prediction. While recent methods demonstrate good forecasting abilities, they have three fundamental limitations. (i). Discrete neural architectures: Interlacing individually parameterized spatial and temporal blocks to encode rich underlying patterns leads to discontinuous latent state trajectories and higher forecasting numerical errors. (ii). High complexity: Discrete approaches complicate models with dedicated designs and redundant parameters, leading to higher computational and memory overheads. (iii). Reliance on graph priors: Relying on predefined static graph structures limits their effectiveness and practicability in real-world applications. In this paper, we address all the above limitations by proposing a continuous model to forecast Multivariate Time series with dynamic Graph neural Ordinary Differential Equations (MTGODE). Specifically, we first abstract multivariate time series into dynamic graphs with time-evolving node features and unknown graph structures. Then, we design and solve a neural ODE to complement missing graph topologies and unify both spatial and temporal message passing, allowing deeper graph propagation and fine-grained temporal information aggregation to characterize stable and precise latent spatial-temporal dynamics. Our experiments demonstrate the superiorities of MTGODE from various perspectives on five time series benchmark datasets. © 2022, CC BY.

关键词： Graph neural networks

来源：评论

学校读者我要写书评

暂无评论

A Density-Dependent Diffusion Model of an Interacting System of Brownian Particles

arXiv

引用

arXiv 2022年

作者： Fouad, Ahmed M. Fouad, Marwa M. Department of Mathematics & Computer Science La Salle University PhiladelphiaPA19141 United States Basic Sciences Department Modern Academy for Engineering & Technology Cairo Egypt

Density-dependent diffusion is a widespread phenomenon in nature. We have examined the density-dependent diffusion behavior of some biological processes such as tumor growth and invasion [23]. Here, we extend our previous work by developing computational techniques to analyze the density-dependent diffusion behavior of one-dimensional interacting particle systems, which have been used to model numerous microscopic processes [17-19], and we apply our techniques to an interacting system of Brownian particles, with hard-core interactions and nearest-neighbor adhesion, known as single-file dynamics. Through large-scale numerical simulations that exploit Monte-Carlo techniques and high-performance computing resources, we show that the diffusion rate in such systems depends on the average particle density. Extensions to the techniques we present here enable researchers to examine the density-dependent diffusion behavior of many physical systems in nature that undergo one-dimensional diffusion associated with a change in the particle density;such as ion transport processes, channeling in zeolites, etc. Copyright © 2022, The Authors. All rights reserved.

关键词： Numerical models

来源：评论

学校读者我要写书评

暂无评论

INFINITE MATRIX PRODUCTS AND HYPERGEOMETRIC ZETA SERIES

arXiv

引用

arXiv 2022年

作者： Wakhare, T. Vignat, C. Department of Electrical Engineering and Computer Science Massachusetts Institute of Technology CambridgeMA United States Department of Mathematics Tulane University New OrleansLA United States

An unpublished identity of Gosper restates a hypergeometric identity for odd zeta values in terms of an infinite product of matrices. We show that this correspondence runs much deeper, and show that many examples of WZ-accelerated series for zeta values lift to infinite matrix products. We also introduce a new matrix subgroup, the Gosper group, which all of our matrix products fall into. Copyright © 2022, The Authors. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Sleep-Like Unsupervised Replay Improves Performance when Data are Limited or Unbalanced

arXiv

引用

arXiv 2024年

作者： Bazhenov, Anthony Dewasurendra, Pahan Krishnan, Giri Delanois, Jean Erik Del Norte High School 16601 Nighthawk Ln San DiegoCA92127 United States Department of Medicine University of California San Diego United States Department of Computer Science & Engineering University of California San Diego 9500 Gilman Drive La Jolla CA92092 United States

The performance of artificial neural networks (ANNs) degrades when training data are limited or imbalanced. In contrast, the human brain can learn quickly from just a few examples. Here, we investigated the role of sleep in improving the performance of ANNs trained with limited data on the MNIST and Fashion MNIST datasets. Sleep was implemented as an unsupervised phase with local Hebbian type learning rules. We found a significant boost in accuracy after the sleep phase for models trained with limited data in the range of 0.5-10% of total MNIST or Fashion MNIST datasets. When more than 10% of the total data was used, sleep alone had a slight negative impact on performance, but this was remedied by finetuning on the original data. This study sheds light on a potential synaptic weight dynamics strategy employed by the brain during sleep to enhance memory performance when training data are limited or imbalanced. © 2024, CC BY-NC-SA.

关键词： Neural networks

来源：评论

学校读者我要写书评

暂无评论

Federated Multimedia Recommendation Systems with Privacy Protection

Federated Multimedia Recommendation Systems with Privacy Pro...

引用

IEEE International Symposium on Broadband Multimedia Systems and Broadcasting (BMSB)

作者： Shih Yu Chang Hsiao-Chun Wu Kun Yan Scott Chih-Hao Huang Yiyan Wu Department of Applied Data Science San Jose State University San Jose CA USA School of Electrical Engineering and Computer Science Louisiana State University Baton Rouge LA USA Guangxi Key Laboratory of Wireless Wideband Communication and Signal Processing Guilin University of Electronic Technology Guilin Guangxi China Department of Information and Telecommunication Guilin University of Electronic Technology Guilin Guangxi China Department of Electrical Engineering National Tsinghua University Hsinchu Taiwan

ISBN: (数字)9798350364262

ISBN: (纸本)9798350364279

In this big data era, distributed machine learning and global data pooling have become quite critical in practical implementation of any multimedia data-science technique. The emerging federated multimedia recommendation systems (FeMRSs) offer a promising approach to address the need of personalized recommendations subject to both users’ privacy and data security. However, there exists no quantitative method for assessing the discrepancy between the large original rating matrix stored on the server and the dimensionality-reduced rating matrix produced on each client’s equipment. Meanwhile, there lacks a systematic means to evaluate the differential privacy (DP), which is a critical security measure in many distributed systems, particularly when sensitive data are processed. In this work, we first introduce a novel personalized dimensionalityreduction algorithm utilizing the matrix sketching technique. This new algorithm effectively control the difference between the original rating matrix on the server and the dimensionality-reduced rating matrix on each client’s equipment. Moreover, a randomized DP matrix factorization algorithm is designed to be executed at each client’s equipment. The theoretical proof is also carried out to show how much DP can be attained by use of the aforementioned randomized DP matrix factorization algorithm. Finally, extensive numerical studies are presented to evaluate the effectiveness of our proposed novel algorithms using both simulated and real datasets for building the FeMRSs.

关键词： Dimensionality reduction Digital multimedia broadcasting Privacy Systematics Distributed databases Filtering algorithms Servers

来源：评论

学校读者我要写书评

暂无评论

Appliancizer: Transforming Web Pages into Electronic Gadgets 33

Appliancizer: Transforming Web Pages into Electronic Gadgets

引用

33rd Annual ACM Symposium on User Interface Software and Technology, UIST 2020

作者： Garza, Jorge Merrill, Devon J. Swanson, Steven Department of Computer Science and Engineering University of California San Diego La JollaCA United States

ISBN: (纸本)9781450375153

Making electronic gadgets that meet today's consumer standards has become a difficult task. Electronic gadgets are expected to have displays with visually appealing interfaces and, at the same time, be physically and screen-interactive, making the development process of these devices time-consuming and challenging. To address this problem, we have created Appliancizer, an online synthesis tool that can automatically generate sophisticated electronic devices from web pages. Appliancizer takes advantage of the similarities between software and physical interfaces to reduce development steps and allow the rapid development of electronic devices. By matching the interface of hardware components with the interface of graphical HTML elements found on web pages, our tool allows a designer to transform HTML elements from a digital to a tangible interface without changing the application source code. Finally, a modular design enables our tool to automatically combine the circuit design and low-level hardware code of selected hardware components into a complete design. Attendees can interact with our online tool and produce manufacturable PCBs from web pages. © 2020 Owner/Author.

关键词： Websites

来源：评论

学校读者我要写书评

暂无评论

An Intelligent System Based on Evolutionary Algorithm for Scheduling university Course Timetable

SSRN

引用

SSRN 2022年

作者： Bimantara, I. Made Satria Yuhana, Umi laili Supriana, I. Wayan Pardede, Eric Department of Informatics Institut Teknologi Sepuluh Nopember 60117 Indonesia Informatics Study Program Faculty of Mathematics and Natural Sciences Udayana University Indonesia Department of Computer Science and Information Technology La Trobe University BundooraVIC3086 Australia

university Course Timetabling (UCT) is a common problem in educational institutions. The preparation of the class schedule must pay attention to available resources without violating set constraints. This research applies Evolutionary Algorithm such as Genetic Algorithm into timetabling system, to replace human expertise in generating optimal lecture schedules. Genetic Algorithm (GA) uses chromosomes to represent the solution of the UCT. One chromosome represents a set of schedule solutions that AG generates. Genes and alleles become components of chromosomes that represent the value of each entity to be optimized. We modeled chromosomes into four gene types, each of which has three allele types, namely day, hour, and room alleles which will be optimized with GA. The N-point and cycle crossovers were implemented into the new chromosome encoding scheme and do not take place at the gene level, but at the allele level by grouping similar allele types before the crossover process is carried out. The success rate is measured by the percentage of the optimal number of schedules generated from the trials. The proposed method provides a success rate of 80 - 85% and can produce an optimal schedule with an average computation time of 8 minutes. GA can reduce human resources and save time in the process of preparing lecture schedules. © 2022, The Authors. All rights reserved.

关键词： Genetic algorithms

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：