检索结果-内蒙古大学图书馆

arXiv 2023年

作者： Wu, Ting Zheng, Rui Gui, Tao Zhang, Qi Huang, Xuanjing School of Computer Science Fudan University China Institute of Modern Languages and Linguistics Fudan University China Shanghai Key Laboratory of Intelligent Information Processing China

Models trained via empirical risk minimization (ERM) are revealed to easily rely on spurious correlations, resulting in poor model generalization. Group distributionally robust optimization (group DRO) can alleviate this problem by minimizing the worst-case loss over pre-defined groups. While promising, in practice factors like expensive annotations and privacy preclude the availability of group labels. More crucially, when taking a closer look at the failure modes of out-of-distribution generalization, the typical procedure of reweighting in group DRO loses efficiency. Hinged on the limitations, in this work, we reformulate the group DRO framework by proposing Q-Diversity. Characterized by an interactive training mode, Q-Diversity relaxes the group identification from annotation into direct parameterization. Furthermore, a novel mixing strategy across groups is presented to diversify the under-represented groups. In a series of experiments on both synthetic and real-world text classification tasks, results demonstrate that Q-Diversity can consistently improve worst-case accuracy under different distributional shifts, outperforming state-of-the-art alternatives. Copyright © 2023, The Authors. All rights reserved.

关键词： Text processing

来源：评论

学校读者我要写书评

暂无评论

Enhancing Contrastive Learning with Noise-Guided Attack: Towards Continual Relation Extraction in the Wild

arXiv

引用

arXiv 2023年

作者： Wu, Ting Liu, Jingyi Zheng, Rui Zhang, Qi Gui, Tao Huang, Xuanjing School of Computer Science Fudan University China Institute of Modern Languages and Linguistics Fudan University China Shanghai Key Laboratory of Intelligent Information Processing China

The principle of continual relation extraction (CRE) involves adapting to emerging novel relations while preserving od knowledge. While current endeavors in CRE succeed in preserving old knowledge, they tend to fail when exposed to contaminated data streams. We assume this is attributed to their reliance on an artificial hypothesis that the data stream has no annotation errors, which hinders real-world applications for CRE. Considering the ubiquity of noisy labels in real-world datasets, in this paper, we formalize a more practical learning scenario, termed as noisy-CRE. Building upon this challenging setting, we develop a noise-resistant contrastive framework named as Noise-guided attack in Contrative Learning (NaCL) to learn incremental corrupted relations. Compared to direct noise discarding or inaccessible noise relabeling, we present modifying the feature space to match the given noisy labels via attacking can better enrich contrastive representations. Extensive empirical validations highlight that NaCL can achieve consistent performance improvements with increasing noise rates, outperforming state-of-the-art baselines. Copyright © 2023, The Authors. All rights reserved.

关键词： Sodium chloride

来源：评论

学校读者我要写书评

暂无评论

Building Agile Workflow Microservice System for HPC Applications Based on Fast-start OSv 27

Building Agile Workflow Microservice System for HPC Applicat...

引用

27th IEEE International Conference on Parallel and Distributed Systems, ICPADS 2021

作者： Hu, Qiang Ma, Lixian Shao, En Tan, Guangming Institute of Computing Technology CAS State Key Laboratory of Computer Architecture Beijing China University of Chinese Academy of Science Beijing China

ISBN: (纸本)9781665408783

The advances of containers have significantly promoted the development of microservice architecture. This architecture splits a monolithic application into multiple independent components and the container orchestrator manages these components by the container in the cloud environment. The feasibility of deploying high performance computing(HPC) applications as microservices has been proven, but the existing container orchestrator incurs a large performance overhead as there is interference between different containers on the same physical host. In this paper, we design an agile workflow microservice system for HPC applications with fast-start OSv. We consider improving HPC workflow performance from two aspects: single OSv startup time optimization and workflow orchestration optimization. For single OSv startup time optimization, we design a fast-start OSv by analyzing the process of OSv startup and finding an optimization by modifying OSv source code. In this way, we get nearly 50% improvement of startup time. For workflow orchestration optimization, we propose four optimization techniques to speed up the execution of workflow by jointly considering OSv and workflow features, namely: node fusion, node merge, image preload, boot delay. Furthermore, we utilize our fast startup OSv to design an orchestration system for efficiently building an agile HPC workflow microservice by Kubevirt. Our experimental results optimization microservice system reduces the execution time by 30% compared with the original deployment with docker. © 2021 IEEE.

关键词： Containers

来源：评论

学校读者我要写书评

暂无评论

RAAMove: A Corpus for Analyzing Moves in Research Article Abstracts 30

RAAMove: A Corpus for Analyzing Moves in Research Article Ab...

引用

Joint 30th International Conference on Computational Linguistics and 14th International Conference on Language Resources and Evaluation, LREC-COLING 2024

作者： Li, Hongzheng Wang, Ruojin Shi, Ge Lv, Xing Lei, Lei Feng, Chong Liu, Fang Lin, Jinkun Mei, Yangguang Xu, Lingnan School of Foreign Languages Key Laboratory of Language Cognition and Computation Ministry of Industry and Information Technology Beijing Institute of Technology 8 Liangxiang East Road Fangshan District Beijing102488 China Faculty of Information Technology Beijing University of Technology 100 Pingleyuan Chaoyang District Beijing100124 China School of Computer Science Beijing Institute of Technology 5 Zhongguancun South St. Haidian District Beijing100081 China

ISBN: (纸本)9782493814104

Move structures have been studied in English for Specific Purposes (ESP) and English for Academic Purposes (EAP) for decades. However, there are few move annotation corpora for Research Article (RA) abstracts. In this paper, we introduce RAAMove, a comprehensive multi-domain corpus dedicated to the annotation of move structures in RA abstracts. The primary objective of RAAMove is to facilitate move analysis and automatic move identification. This paper provides a thorough discussion of the corpus construction process, including the scheme, data collection, annotation guidelines, and annotation procedures. The corpus is constructed through two stages: initially, expert annotators manually annotate high-quality data;subsequently, based on the human-annotated data, a BERT-based model is employed for automatic annotation with the help of experts' modification. The result is a large-scale and high-quality corpus comprising 33,988 annotated instances. We also conduct preliminary move identification experiments using the BERT-based model to verify the effectiveness of the proposed corpus and model. The annotated corpus is available for academic research purposes and can serve as essential resources for move analysis, English language teaching and writing, as well as move/discourse-related tasks in Natural Language Processing (NLP). © 2024 ELRA Language Resource Association: CC BY-NC 4.0.

关键词： Abstracting

来源：评论

学校读者我要写书评

暂无评论

Large Scale Model Enabled Semantic Communications Based on Robust Knowledge Distillation

Large Scale Model Enabled Semantic Communications Based on R...

引用

2024 IEEE Global Communications Conference, GLOBECOM 2024

作者： Ding, Kuiyuan Liu, Fangfang Yang, Yang Chen, Mingzhe Guo, Caili Beijing University of Posts and Telecommunications Beijing Key Laboratory of Network System Architecture and Convergence School of Information and Communication Engineering Beijing100876 China University of Miami Department of Electrical and Computer Engineering The Institute for Data Science and Computing Coral GablesFL33146 United States Beijing University of Posts and Telecommunications Beijing Laboratory of Advanced Information Networks School of Information and Communication Engineering Beijing100876 China

ISBN: (纸本)9798350351255

Large scale artificial intelligence (AI) models possess excellent capabilities in semantic representation and understanding, making them particularly well-suited for semantic encoding and decoding. However, the substantial scale of these AI models imposes unacceptable computational resources and communication delays. To address this issue, we propose a semantic communication scheme based on robust knowledge distillation (RKD-SC) for large scale model enabled semantic communications. In the considered system, a transmitter extracts the features of the source image for robust transmission and accurate image classification at the receiver. To effectively utilize the superior capability of large scale model while make the cost affordable, we first transfer knowledge from a large scale model to a smaller scale model to serve as the semantic encoder. Then, to enhance the robustness of the system against channel noise, we propose a channel-aware autoencoder (CAA) based on the Transformer architecture. Experimental results show that the encoder of proposed RKD-SC system can achieve over 93.3% of the performance of a large scale model while compressing 96.67% number of parameters. Code: https://***/echojayne/RKD-SC. © 2024 IEEE.

关键词： Channel coding

来源：评论

学校读者我要写书评

暂无评论

Predicting Large Language Model Capabilities on Closed-Book QA Tasks Using Only Information Available Prior to Training

arXiv

引用

arXiv 2025年

作者： Jiang, Changhao Zhang, Ming Ye, Junjie Fan, Xiaoran Cao, Yifei Sun, Jiajun Xi, Zhiheng Dou, Shihan Dong, Yi Shen, Yujiong Tong, Jingqi Wang, Zhen Liang, Tao Fei, Zhihui Wan, Mingyang Ma, Guojun Zhang, Qi Gui, Tao Huang, Xuanjing School of Computer Science Fudan University Shanghai China ByteDance China Shanghai Key Laboratory of Intelligent Information Processing Fudan University Shanghai China Institute of Modern Languages and Linguistics Fudan University Shanghai China

The GPT-4 technical report from OpenAI suggests that model performance on specific tasks can be predicted prior to training, though methodologies remain unspecified. This approach is crucial for optimizing resource allocation and ensuring data alignment with target tasks. To achieve this vision, we focus on predicting performance on Closed-book Question Answering (CBQA) tasks, which are closely tied to pre-training data and knowledge retention. We address three major challenges: 1) mastering the entire pre-training process, especially data construction;2) evaluating a model’s knowledge retention;and 3) predicting task-specific knowledge retention using only information available prior to training. To tackle these challenges, we pre-train three large language models (i.e., 1.6B, 7B, and 13B) using 560k dollars and 520k GPU hours. We analyze the pre-training data with knowledge triples and assess knowledge retention using established methods. Additionally, we introduce the SMI metric, an information-theoretic measure that quantifies the relationship between pre-training data, model size, and task-specific knowledge retention. Our experiments reveal a strong linear correlation (R2 > 0.84) between the SMI metric and the model’s accuracy on CBQA tasks across models of varying sizes (i.e., 1.1B, 1.6B, 7B, and 13B). The dataset, model, and code are available at https://***/yuhui1038/SMI. Copyright © 2025, The Authors. All rights reserved.

关键词： Resource allocation

来源：评论

学校读者我要写书评

暂无评论

Analysis and Design of Mobile Model for DTN

引用

Chinese Journal of Electronics 2023年第4期18卷 635-638页

作者： Zhiting Lin Yugui Qu Yusheng Ji Baohua Zhao Department of Computer Science University of Science and Technology of China Hefei China Province Key Laboratory of Software in Computing and Communication Hefei China Information Systems Architecture Science Research Division National Institute of Informatics Tokyo Japan

Delay tolerant networks may become unexpectedly partitioned due to node mobility or variation in signal strength. However, most widely used models in some relative works are generally very simplistic. In order to exploit intelligent forwarding algorithms, a novel nodal mobile model of delay tolerant networks is presented to map reality with more accuracy. And several approaches are introduced to analyze the network structure, such as n-cliques, n-clans, degree, closeness and betweenness. Our research results showed that the centralizations became smaller when the wireless connections were concerned. It meant that quite a number of nodes became potential relays because of the new structure of networks.

关键词： Wireless communication Analytical models Wireless sensor networks Protocols Social networking (online) Sensitivity analysis Social network theory

来源：评论

学校读者我要写书评

暂无评论

ToolEyes: Fine-Grained Evaluation for Tool Learning Capabilities of Large Language Models in Real-world Scenarios

arXiv

引用

arXiv 2024年

作者： Ye, Junjie Li, Guanyu Gao, Songyang Huang, Caishuang Wu, Yilong Li, Sixian Fan, Xiaoran Dou, Shihan Ji, Tao Zhang, Qi Gui, Tao Huang, Xuanjing School of Computer Science Fudan University China Institute of Modern Languages and Linguistics Fudan University China Key Laboratory of Intelligent Information Processing Fudan University Shanghai China Research Institute of Intelligent Complex Systems Fudan University China Pengcheng Laboratory China

Existing evaluations of tool learning primarily focus on validating the alignment of selected tools (e.g., various APIs) for large language models (LLMs) with expected outcomes. However, these approaches rely on a limited set of scenarios where answers can be pre-determined. Furthermore, a sole emphasis on outcomes disregards the complex capabilities required for LLMs to effectively use tools. To tackle this issue, we propose ToolEyes, a fine-grained system tailored for the evaluation of the LLMs’ tool learning capabilities in authentic scenarios. The system meticulously examines seven real-world scenarios, analyzing five dimensions crucial to LLMs in tool learning: format alignment, intent comprehension, behavior planning, tool selection, and answer organization. Additionally, ToolEyes incorporates a tool library boasting approximately 600 tools, serving as an intermediary between LLMs and the physical world. Evaluations involving ten LLMs across three categories reveal a preference for specific scenarios and limited cognitive abilities in tool learning. Intriguingly, expanding the model size even exacerbates the hindrance to tool learning. The code and data are available at https://***/Junjie-Ye/ToolEyes. Copyright © 2024, The Authors. All rights reserved.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Safety Guaranteed Power-Delivered-to-Load Maximization for Magnetic Wireless Power Transfer

Safety Guaranteed Power-Delivered-to-Load Maximization for M...

引用

IEEE Annual Joint Conference: INFOCOM, IEEE computer and Communications Societies

作者： Wangqiu Zhou Xinyu Wang Hao Zhou Shenyao Jiang Zhi Liu Yusheng Ji LINKE Lab School of Computer Science and Technology University of Science and Technology of China CAS Key Laboratory of Wireless-Optical Communications University of Science and Technology of China Deqing Alpha Innovation Institute Huzhou China Department of Computer and Network Engineering The University of Electro-Communications Tokyo Japan Information Systems Architecture Science Research Division National Institute of Informatics Tokyo Japan

ISBN: (数字)9798350383508

ISBN: (纸本)9798350383515

Electromagnetic radiation (EMR) safety has always been a critical reason for hindering the development of magneticenabled wireless power transfer technology. People focus on the actual received energy at charging devices while paying attention to their health. Thus, we study this significant problem in this paper, and propose a universal safety guaranteed power-delivered-to-load (PDL) maximization scheme (called SafeGuard). Technically, we first utilize the off-the-shelf electromagnetic simulator to perform the EMR distribution analysis to ensure the universality of the method. Then, we innovatively introduce the concept of multiple importance sampling for achieving efficient EMR safety constraint extraction. Finally, we treat the proposed optimization problem as an optimal boundary point search problem from the perspective of space geometry, and devise a brand-new grid-based multi-constraint parallel processing algorithm to efficiently solve it. We implement a system prototype for SafeGuard, and conduct extensive experiments to evaluate it. The results indicate that our SafeGuard can obviously improve the achieved PDL by up to 1.75× compared with the state-of-the-art baseline while guaranteeing EMR safety. Furthermore, SafeGuard can accelerate the solution process by 29.12× compared with the traditional numerical method to satisfy the fast optimization requirement of wireless charging systems.

关键词： Performance evaluation Monte Carlo methods System performance Inductive charging Prototypes Wireless power transfer Parallel processing

来源：评论

学校读者我要写书评

暂无评论

Energy-Efficient and Reliable Federated Learning in Heterogeneous Mobile-Edge Computing

Energy-Efficient and Reliable Federated Learning in Heteroge...

引用

Symposia and Workshops on Ubiquitous, Autonomic and Trusted Computing, UIC-ATC

作者： Yan Xu Liying Li Peijin Cong Junlong Zhou School of Computer Science and Engineering Nanjing University of Science and Technology China State Key Laboratory of Computer Architecture Institute of Computing Technology Chinese Academy of Sciences China

As a decentralized machine learning paradigm, Federated Learning (FL) is an emerging technique to protect user data privacy in Mobile Edge Computing (MEC). FL adopts the idea of distributed privacy computing, enabling the terminal devices to train machine learning models required by servers locally and upload model parameters to servers for aggregation. However, in the process of coupling FL with MEC systems, device heterogeneity, malicious use of poisoned data by users and limited system resources will seriously affect FL training accuracy, system reliability and energy efficiency. In this paper, a highly reliable and energy-efficient FL scheme is designed to solve these issues. Specifically, we first propose a calculation method for estimating the reliability of heterogeneous devices. Then, we design a highly reliable device selection scheme based on devices’ reliability, computing power, and participation times. Finally, we utilize the slack time in FL training to dynamically adjust the voltage and frequency of devices for reducing the energy consumption of training. Experimental results show that our proposed scheme improves the accuracy by up to 36.45% on average and saves energy by up to 48.63% when compared to the baseline and benchmarking methods.

关键词：

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：