检索结果-内蒙古大学图书馆

1st Workshop on Language Models for Low-Resource Languages, LoResLM 2025 - co-located at the 31st International Conference on Computational Linguistics, COLING 2025

作者： Lasheras, Uriel Pinheiro, Vládia Postgraduate Program in Applied Informatics University of Fortaleza Ceará Brazil

ISBN: (纸本)9798891762152

Large Language Models (LLMs) are increasingly central to the development of generative AI across diverse fields. While some anticipate these models may mark a step toward artificial general intelligence, their ability to handle complex causal reasoning remains unproven. Causal reasoning is essential for true general intelligence, particularly at Pearl’s interventional and counterfactual levels. In this work, we introduce ***, a dataset of over 8,000 natural causal questions in Portuguese, collected from real human interactions. Built upon a novel three-axis taxonomy, *** categorizes questions by causal intent, action requirements, and the level of causal reasoning needed (associational, interventional, or counterfactual). Our findings from evaluating ***’s seed questions with GPT-4o reveal that this LLM faces challenges in handling interventional and relation-seeking causal queries. These results suggest limitations in using GPT-4o for extending causal question annotations and highlight the need for improved LLM strategies in causal reasoning. *** provides a foundation for advancing LLM capabilities in causal understanding, particularly for the Portuguese-speaking world. © 2025 Association for Computational Linguistics.

关键词： Computational linguistics

来源：评论

学校读者我要写书评

暂无评论

Design and Evaluation of an Event-Driven Cloud-Based Architecture for a Remote Patient Monitoring System

Design and Evaluation of an Event-Driven Cloud-Based Archite...

引用

IEEE International Conference on Software Architecture Workshops (ICSAW)

作者： Pedro Fiuza Linhares Pedro H. Gaya Wanderley Marza de Sousa Zaranza Maria Andréia Formico Rodrigues Nabor C. Mendonça Post Graduate Program in Applied Informatics University of Fortaleza Fortaleza CE Brazil Center of Technological Sciences University of Fortaleza Fortaleza CE Brazil Intensive Care Unit Hospital Instituto Dr. José Frota (IJF) Fortaleza CE Brazil Post Graduate Program in Applied Informatics/GIRA Lab University of Fortaleza Fortaleza CE Brazil

ISBN: (数字)9798331533366

ISBN: (纸本)9798331533373

This paper presents an event-driven cloud-based serverless architecture for a remote patient monitoring system. The architecture is designed to be scalable, resilient, and cost-effective, and leverages managed cloud services to handle load variations and service faults. To evaluate the architecture, we conducted comprehensive capacity, scalability, and resilience tests. The test results show that the architecture can meet the performance, scalability, and reliability requirements of a remote patient monitoring system. Additionally, a detailed analysis of operational costs in the cloud confirms the cost-effectiveness of the solution and highlights practical opportunities for improvements.

关键词： Cloud computing Patient monitoring Software architecture Scalability Computer architecture Load management Performance analysis Reliability Resilience Testing

来源：评论

学校读者我要写书评

暂无评论

Principal Curvatures Estimation with Applications to Single Cell Data

Principal Curvatures Estimation with Applications to Single ...

引用

2025 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2025

作者： Zhang, Yanlei Mezrag, Lydia Sun, Xingzhi Xu, Charles Macdonald, Kincaid Bhaskar, Dhananjay Krishnaswamy, Smita Wolf, Guy Rieck, Bastian Université de Montréal Dept. of Math. & Stat Canada Mila - Quebec AI Institute MontréalQC Canada Yale University Dept. of Comp. Sci. United States Dept. of Genetics United States Applied Mathematics Program New HavenCT United States Université de Fribourg Department of Informatics FR Fribourg Switzerland Helmholtz Zentrum München Institute of AI for Health BY Munich Germany

ISBN: (纸本)9798350368741

The rapidly growing field of single-cell transcriptomic sequencing (scRNAseq) presents challenges for data analysis due to its massive datasets. A common method in manifold learning consists in hypothesizing that datasets lie on a lower dimensional manifold. This allows to study the geometry of point clouds by extracting meaningful descriptors like curvature. In this work, we will present Adaptive Local PCA (AdaL-PCA), a data-driven method for accurately estimating various notions of intrinsic curvature on data manifolds, in particular principal curvatures for surfaces. The model relies on local PCA to estimate the tangent spaces. The evaluation of AdaL-PCA on sampled surfaces shows state-of-the-art results. Combined with a PHATE embedding, the model applied to single-cell RNA sequencing data allows us to identify key variations in the cellular differentiation. © 2025 IEEE.

关键词： Gaussian curvature Principal curvature principal directions single-cell

来源：评论

学校读者我要写书评

暂无评论

A Reconfigurable Compute-In-Memory Design with Adaptive Bit-Precision and Power Adjustment

A Reconfigurable Compute-In-Memory Design with Adaptive Bit-...

引用

Consumer Technology (ICCT-Pacific), International Conference on

作者： I-Chieh Hsu Chun-Fu Chen Dao-Long Huang Yun-Kai Yeh Wei-Chieh Lee Qi-Hua Lin Jinjia Zhou I-Chyn Wey Electrical Engineering Department College of Engineering Chang-Gung University Taiwan School of Electrical Engineering National Taiwan University of Science and Technology Taiwan Master Program in Nano-Electronic Engineering and Design Chang Gung University Taiwan Department of Applied Informatics Faculty of Science and Engineering Hosei University Japan

ISBN: (数字)9798331504120

ISBN: (纸本)9798331504137

With the successful development of artificial intelligence (AI), Convolutional Neural Networks (CNNs) occupy a large amount of computing time and power consumption in the entire AI computing process. However, in traditional von Neumann architectures, the separation of the computing element and memory leads to the memory-wall problem of memory data transmission bandwidth from memory to processing element when executing multiply-accumulate-based CNN operations. The data transmission time and power consumption all are much higher than the CNN computing part. Therefore, this paper proposes a memory-in-computation design that can flexibly adjust energy usage, achieve high energy efficiency, and support multiple operation frequencies. By controlling the switches of each memory row, unnecessary energy consumption is avoided, leading to a reduction in total power consumption by 15% to 60%. Additionally, by adjusting the pulse width, the charging power consumption of capacitors is reduced by 65% per charge cycle.

关键词： Energy consumption Power demand Capacitors Bandwidth In-memory computing Control systems Energy efficiency Convolutional neural networks Data communication Artificial intelligence

来源：评论

学校读者我要写书评

暂无评论

Investigating the computation of smoothed deceptive routes in topographic terrains for agent-based simulation systems

Investigating the computation of smoothed deceptive routes i...

引用

Computational Intelligence in Security, Defence and Biometrics (CISDB), IEEE Symposium on

作者： Luigi P. Souza Thiago R. S. Leão Crhistopher Lenhard Edison P. Freitas Luis A. L. Silva Graduate Program in Computer Science Federal Univ. of Santa Maria Santa Maria Brazil Undergraduate Course in Computer Science Federal Univ. of Santa Maria Santa Maria Brazil Institute of Informatics Federal Univ. of Rio Grande do Sul Porto Alegre Brazil Applied Computing Department Federal Univ. of Santa Maria Santa Maria Brazil

ISBN: (数字)9798331508296

ISBN: (纸本)9798331508302

The capacity to deceptively move in hilly terrains is fundamental to agents in simulation systems for tactical and strategic military training. Such an ability to deceive the adversary can ensure a relevant advantage by hiding the real goals of a mission. Using pairs of real and deceptive mission goals, this paper investigates the planning of realistic deceptive routes for simulated agents. With real-world terrain elevation maps, it shows how to explore pathfinding algorithms with relevant path-smoothing characteristics (Theta* and WJPS*, contrasting with the standard $A^{*}$ algorithm) in terrains with pronounced relief features. The study analyzes the effects of terrain elevation costs and the representation of relief contour lines on the determination of more realistic deceptive paths. This work also investigates how users can adjust a Last Topographic Deceptive Point ($L D P_{T}$) calculation to enhance the pathfinding algorithm’s ability to produce more deceptively dense and topographically aware routes. Experimental results for different deceptive topographic path planning strategies are evaluated according to statistical models showing that Theta*, despite being slower than the base $A^{*}$ method in most cases, generated smoothed paths while maintaining a similar deception density for the proposed strategies. On the other hand, WJPS outperformed both in execution time for certain strategies while maintaining the smoothed path characteristic and resulting in a path with lower deceptive capacity.

关键词： Training Hands Biometrics Costs Path planning Planning Security Standards Computational intelligence

来源：评论

学校读者我要写书评

暂无评论

Drought propagation in an Andean basin in South America: a spatio-temporal analysis approach with monitoring indices

引用

Arabian Journal of Geosciences 2025年第3期18卷 1-21页

作者： Ocampo-Marulanda, Camilo Mera-Franco, Julián David Lizcano, Víctor Augusto Carvajal-Escobar, Yesid Canchala, Teresita Zorrilla-Millan, Daniel Felipe Caicedo-Bravo, Eduardo Water Resources Engineering and Soil Research Group (IREHISA) School of Natural Resources and Environmental Engineering Universidad del Valle Cali Colombia Graduate Program in Biometrics and Applied Statistics Department of Statistics and Informatics Federal Rural University of Pernambuco Recife Brazil School of Exact Sciences and Engineering Universidad Sergio Arboleda Bogotá Colombia Programa de Ingeniería Ambiental Facultad de Ingeniería Universidad Mariana Pasto Colombia School of Electrical and Electronic Engineering Universidad del Valle Cali Colombia

Drought is a climatic phenomenon that has increased in severity over time. This study analyzed drought spread in the Alto Patía basin (13,047 km2) in South America, applying the Standardized Precipitation Index (SPI), Vegetation Health Index (VHI), and Streamflow Drought Index (SDI) to assess meteorological, agricultural, and hydrological drought, respectively. Synchronous and asynchronous correlations were estimated, and Wavelet analysis was conducted to verify relationships among drought types. Results show that the lag between drought types varies based on the intensity, temporal continuity, and spatial extent of meteorological drought. A lag of zero months (1–2 months) was observed between meteorological and hydrological drought when moderate (severe) meteorological drought partially (fully) covers the basin. Agricultural drought, however, does not always correspond to meteorological drought during isolated rainfall events. As the first study in Colombia to systematically connect these drought types, this research addresses a gap in understanding how drought impacts progress across systems in regions with complex rainfall patterns. These findings offer critical insights into the interdependencies among drought types, supporting enhanced drought monitoring and early warning systems in similar climatic regions. By identifying specific lags between meteorological and hydrological droughts, this study provides practical guidance for land management, water conservation, and crop planning, offering a foundation for future research on drought resilience strategies.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Principal Curvatures Estimation with Applications to Single Cell Data

Principal Curvatures Estimation with Applications to Single ...

引用

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

作者： Yanlei Zhang Lydia Mezrag Xingzhi Sun Charles Xu Kincaid Macdonald Dhananjay Bhaskar Smita Krishnaswamy Guy Wolf Bastian Rieck Dept. of Math. & Stat. Université de Montréal Mila – Quebec AI Institute Montréal CA Dept. of Comp. Sci. Yale University Dept. of Genetics Applied Mathematics Program New Haven CT USA Helmholtz Zentrum München Institute of AI for Health Munich BY DE Department of Informatics Université de Fribourg Fribourg CH

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

关键词： Manifolds Sequential analysis Adaptation models Data analysis RNA Transcriptomics Curve fitting Speech processing Tuning Principal component analysis

来源：评论

学校读者我要写书评

暂无评论

Telling stories about system use: Capturing collective tacit knowledge for system maintenance

Telling stories about system use: Capturing collective tacit...

引用

19th International Conference on Software Engineering and Knowledge Engineering, SEKE 2007

作者： De Oliveira, Adriana Cristina De Araujo, Renata Mendes Borges, Marcos R.S. Program in Informatics IM and NCE UFRJ Brazil Department of Applied Informatics UNIRIO Brazil

ISBN: (纸本)9781627486613

One of the great system engineering challenges has been how to understand user needs. The evolution of software requirements methodologies and techniques has been motivated by the belief that understanding business, client and user needs is a key activity for successfull system deployment and use. During maintenance, system users are again the most suitable actors able to recognize special kind of information about system use, usually hidden from the system maintenance team. Benefits and problems of using the system in ther daily activities, opinions, suggestions of improvement, beliefs, 'shortcuts' for greater productivity while using the system etc, usually remain as user tacit knowledge. This article presents the use of Group Storytelling as a knowledge capture and sharing technique, intending to use the colaborative creation of stories about the use of information systems as a way to assist software maintenance team activities. Copyright © (2007) by Knowledge Systems Institute (KSI).

关键词： Computer software maintenance

来源：评论

学校读者我要写书评

暂无评论

μ-brkga: A parallel biased random-key genetic algorithm applied to nesting problems 47

μ-brkga: A parallel biased random-key genetic algorithm app...

引用

47th International Conference on Computers and Industrial Engineering: How Digital Platforms and Industrial Engineering are Transforming Industry and Services, CIE 2017

作者： Amaro, Bonfim Pinheiro, Plácido Rogério Graduate Program in Applied Informatics University of Fortaleza Fortaleza Brazil

The Nesting Problems are a special class of cutting and packing (C&P) problems in which a set of items with arbitrary geometry must be positioned in a surface with a non-fixed length, aiming a minimization of the require area for allocation of all given demand. It is commonly applied in industries that operate in various types of manufacturers and exporters activities (e.g., shipbuilding, clothes, and glass). In this work, we propose a parallel biased random-key genetic algorithm with multiple populations (μ-BRKGA) to the Nesting Problems using a collision-free region (CFR) concept as position method, intending at a more efficient and fastest layout compaction to get "good" solutions. The problem layout of the proposed algorithm, a solution, is represented for placement order in the surface and the corresponding orientation. As the way of to evaluate our (μ-BRKGA) algorithm, computational tests using benchmark instances were applied, discussed and compared with different existing literature methods.

关键词： Genetic algorithms

来源：评论

学校读者我要写书评

暂无评论

A Methodology for Risk Management of Generative AI based Systems 32

A Methodology for Risk Management of Generative AI based Sys...

引用

32nd International Conference on Software, Telecommunications and Computer Networks, SoftCOM 2024

作者： Filho, Raimir Holanda Colares, Daniel University of Fortaleza Postgraduate Program in Applied Informatics Fortaleza Brazil

ISBN: (纸本)9789532901382

This paper addresses cybersecurity challenges in integrating Generative Artificial Intelligence (GenAI) into applications. It proposes a risk analysis approach that focuses on identifying threats and vulnerabilities exploited by malicious actors, which is validated through a case study on prompt injection attacks. The results demonstrated the efficacy of our approach in identifying and quantifying vulnerabilities in Large Language Models (LLMs). © 2024 University of Split, FESB.

关键词： Cyber Security Generative AI Large Language Models Risk Management

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：