检索结果-内蒙古大学图书馆

arXiv 2022年

作者： Zhu, Jianggang Wang, Zheng Chen, Jingjing Chen, Yi-Ping Phoebe Jiang, Yu-Gang Shanghai Key Lab of Intelligent Information Processing School of Computer Science Fudan University China Shanghai Collaborative Innovation Center on Intelligent Visual Computing China Department of Computer Science and Information Technology La Trobe University Australia

Real-world data typically follow a long-tailed distribution, where a few majority categories occupy most of the data while most minority categories contain a limited number of samples. Classification models minimizing cross-entropy struggle to represent and classify the tail classes. Although the problem of learning unbiased classifiers has been well studied, methods for representing imbalanced data are under-explored. In this paper, we focus on representation learning for imbalanced data. Recently, supervised contrastive learning has shown promising performance on balanced data recently. However, through our theoretical analysis, we find that for long-tailed data, it fails to form a regular simplex which is an ideal geometric configuration for representation learning. To correct the optimization behavior of SCL and further improve the performance of long-tailed visual recognition, we propose a novel loss for balanced contrastive learning (BCL). Compared with SCL, we have two improvements in BCL: class-averaging, which balances the gradient contribution of negative classes;class-complement, which allows all classes to appear in every mini-batch. The proposed balanced contrastive learning (BCL) method satisfies the condition of forming a regular simplex and assists the optimization of cross-entropy. Equipped with BCL, the proposed two-branch framework can obtain a stronger feature representation and achieve competitive performance on long-tailed benchmark datasets such as CIFAR-10-LT, CIFAR-100-LT, ImageNet-LT, and iNaturalist2018. Our code is available at this URL. Copyright © 2022, The Authors. All rights reserved.

关键词： Benchmarking

来源：评论

学校读者我要写书评

暂无评论

HPC-Coder: Modeling Parallel Programs using Large Language Models

HPC-Coder: Modeling Parallel Programs using Large Language M...

引用

ISC High Performance 2024 Research Paper Proceedings (39th International Conference)

作者： Daniel Nichols Aniruddha Marathe Harshitha Menon Todd Gamblin Abhinav Bhatele Department of Computer Science University of Maryland College Park MD USA Center for Applied Scientific Computing Lawrence Livermore National Laboratory Livermore CA USA Livermore Computing Lawrence Livermore National Laboratory Livermore CA USA

Parallel programs in high performance computing (HPC) continue to grow in complexity and scale in the exascale era. The diversity in hardware and parallel programming models make developing, optimizing, and maintaining parallel software even more burdensome for developers. One way to alleviate some of these burdens is with automated development and analysis tools. Such tools can perform complex and/or remedial tasks for developers that increase their productivity and decrease the chance for error. Until recently, such tools for code development and performance analysis have been limited in the complexity of tasks they can perform, especially for parallel programs. However, with recent advancements in language modeling, and the availability of large amounts of open-source code related data, these tools have started to utilize predictive language models to automate more complex tasks. In this paper, we show how large language models (LLMs) can be applied to tasks specific to high performance and scientific codes. We introduce a new dataset of HPC and scientific codes and use it to fine-tune several pre-trained models. We compare several pre-trained LLMs on HPC-related tasks and introduce a new model, HPC-Coder, fine-tuned on parallel codes. In our experiments, we show that this model can auto-complete HPC functions where generic models cannot, decorate for loops with OpenMP pragmas, and model performance changes in scientific application repositories as well as programming competition solutions.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Alleviating Shifted Distribution in Human Preference Alignment through Meta-Learning 39

Alleviating Shifted Distribution in Human Preference Alignme...

引用

39th Annual AAAI Conference on Artificial Intelligence, AAAI 2025

作者： Dou, Shihan Liu, Yan Zhou, Enyu Gao, Songyang Li, Tianlong Xiong, Limao Zhao, Xin Jia, Haoxiang Ye, Junjie Zheng, Rui Gui, Tao Zhang, Qi Huang, Xuanjing School of Computer Science Fudan University Shanghai China Ant Group Shanghai China School of Computer Science Peking University Beijing China Institute of Modern Languages and Linguistics Fudan University Shanghai China Key Laboratory of Intelligent Information Processing Fudan University Shanghai China Shanghai Collaborative Innovation Center of Intelligent Visual Computing Shanghai China

ISBN: (纸本)157735897X

The capability of the reward model (RM) is crucial for the success of Reinforcement Learning from Human Feedback (RLHF) in aligning with human preferences. However, as training progresses, the output space distribution of the policy model shifts. The RM, initially trained on responses sampled from the output distribution of the early policy model, gradually loses its ability to distinguish between responses from the newly shifted distribution. This issue is further compounded when the RM, trained on a specific data distribution, struggles to generalize to examples outside of that distribution. These two issues can be united as a challenge posed by the shifted distribution of the environment. To surmount this challenge, we introduce MetaRM, a novel method leveraging meta-learning to adapt the RM to the shifted environment distribution. MetaRM optimizes the RM in an alternating way, by preserving both the preferences of the original preference pairs, as well as maximizing discrimination power over new examples of the shifted distribution. Extensive experiments demonstrate that MetaRM can iteratively enhance the performance of human preference alignment by improving the RM’s capacity to identify subtle differences in samples of shifted distributions. Copyright © 2025, Association for the Advancement of Artificial Intelligence (***). All rights reserved.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Silver Syntax Pre-training for Cross-Domain Relation Extraction

arXiv

引用

arXiv 2023年

作者： Bassignana, Elisa Ginter, Filip Pyysalo, Sampo van der Goot, Rob Plank, Barbara Department of Computer Science IT University of Copenhagen Denmark TurkuNLP Department of Computing University of Turku Finland MaiNLP Center for Information and Language Processing LMU Munich Germany

Relation Extraction (RE) remains a challenging task, especially when considering realistic out-of-domain evaluations. One of the main reasons for this is the limited training size of current RE datasets: obtaining high-quality (manually annotated) data is extremely expensive and cannot realistically be repeated for each new domain. An intermediate training step on data from related tasks has shown to be beneficial across many NLP tasks. However, this setup still requires supplementary annotated data, which is often not available. In this paper, we investigate intermediate pre-training specifically for RE. We exploit the affinity between syntactic structure and semantic RE, and identify the syntactic relations which are closely related to RE by being on the shortest dependency path between two entities. We then take advantage of the high accuracy of current syntactic parsers in order to automatically obtain large amounts of low-cost pre-training data. By pre-training our RE model on the relevant syntactic relations, we are able to outperform the baseline in five out of six cross-domain setups, without any additional annotated data. © 2023, CC BY.

关键词： Syntactics

来源：评论

学校读者我要写书评

暂无评论

Distributed optimization for IoT attack detection using federated learning and Siberian Tiger optimizer

引用

ICT Express 2025年第3期11卷 542-546页

作者： Gupta, Brij B. Gaurav, Akshat Alhalabi, Wadee Arya, Varsha Alharbi, Eman Chui, Kwok Tai Department of Computer Science and Information Engineering Asia University Taichung 413 Taiwan Department of Medical Research China Medical University Hospital China Medical University Taichung Taiwan Symbiosis Centre for Information Technology (SCIT) Symbiosis International University Pune India School of Cybersecurity Korea University Seoul South Korea Kyung Hee University 26 Kyungheedae-ro Dongdaemun-gu Seoul South Korea Ronin Institute Montclair NJ United States International Center for AI and Cyber Security Research and Innovations (CCRI) Asia University Taichung 413 Taiwan Department of Computer Science Immersive Virtual Reality Research Group King Abdulaziz University Jeddah Saudi Arabia Hong Kong Metropolitan University Hong Kong SAR China Center for Interdisciplinary Research University of Petroleum and Energy Studies (UPES) Dehradun India Department of Computer Science Faculty of Computing and Information Technology King Abdulaziz University Jeddah Saudi Arabia

The rapid growth of IoT devices has heightened the risk of botnet attacks, calling for scalable and distributed detection solutions. In this context, this study proposes a distributed optimization system for IoT attack detection using CNN model utilizing federated learning. After optimizing the hyperparameters of the model at the server, the Siberian Tiger Optimization (STO) method distributes these values to clients for dispersed training. Our model achieves accuracy, recall, and precision of 0.89978, 0.94355, and 0.94455, respectively, using the N-BaIoT dataset. These findings show, in spite of latency issues, the efficiency of federated learning in distributed IoT security systems. © 2025

关键词： CNN model Distributed optimization Federated learning IoT attack detection Siberian Tiger Optimization (STO)

来源：评论

学校读者我要写书评

暂无评论

Eye Drop Instillation Success and Hand Function in Adults with Glaucoma: A Pilot Study

引用

Ophthalmology Glaucoma 2025年第3期8卷 275-284页

作者： Weber, Madeline K. Kabil, Gül G. Niziol, Leslie M. Urrego, Daniel Duque Haire, Cameron Cain, Stephen M. Sample, Alanson P. Burke, David T. Brown, Susan H. Newman-Casey, Paula Anne University of Michigan Medical School Ann Arbor MI United States University of Michigan School of Kinesiology Ann Arbor MI United States Department of Ophthalmology and Visual Sciences University of Michigan W.K. Kellogg Eye Center Ann Arbor MI United States Department of Chemical and Biomedical Engineering West Virginia University Morgantown WV United States Department of Electrical Engineering and Computer Science University of Michigan Ann Arbor MI United States Department of Human Genetics University of Michigan Ann Arbor MI United States

Purpose: To investigate hand function and eye drop instillation success in adults with and without glaucoma. Design: Cross-sectional pilot study. Subjects: Adults aged ≥ 65 years with glaucoma who use eye drops daily and adults aged 65+ without glaucoma who do not regularly use eye drops. Methods: Hand function was evaluated using the Saehan Hydraulic Pinch Gauge, Jamar Hand Dynamometer, Grooved Pegboard Test (GPT), Arthritis Hand Function Tests (AHFT), Semmes-Weinstein Monofilament Test, and a tactile pattern recognition device. Eye drop instillation success was analyzed through videography. Main Outcome Measures: Proportion of successful eye drop instillation trials assessed by (1) overall success in getting at least one drop in the eye;(2) average number of drops dispensed;and (3) bottle tip contact with the eye, eyelashes, or skin. Hand function measures: mean pinch and grip strength, GPT and AHFT completion times, smallest detected monofilament, tactile pattern identification time, and accuracy. Results: A total of 25 participants with glaucoma and 79 participants without glaucoma were included. There was no difference in the proportion of trials where at least one drop was successfully instilled, disregarding bottle tip contact (glaucoma: 0.95, nonglaucoma: 0.91;P = 0.88). Participants with glaucoma were more likely to make contact between the bottle tip and the eye, eyelashes, or skin compared with those without (glaucoma: 0.49, nonglaucoma: 0.28;P = 0.01). Participants with glaucoma dispensed a similar number of drops as those without (glaucoma: 1.37, nonglaucoma: 1.46;P = 0.47). Participants with glaucoma had significantly reduced pinch strength (glaucoma: 4.8 kg, nonglaucoma 6.1 kg, P = 0.01), grip strength (glaucoma: 23.4 kg, nonglaucoma: 27.7 kg, P = 0.02), longer completion times for the GPT (glaucoma: 113.5 seconds, nonglaucoma: 85.5 seconds, P = 0.02), and specific AHFT tasks of fastening/unfastening buttons (glaucoma: 36.6 seconds, nonglaucoma: 27.7

关键词： Eye drop instillation success Glaucoma Hand function Sensorimotor deficits

来源：评论

学校读者我要写书评

暂无评论

Direct GPU Compilation and Execution for Host Applications with OpenMP Parallelism 8

Direct GPU Compilation and Execution for Host Applications w...

引用

8th IEEE/ACM Annual Workshop on the LLVM Compiler Infrastructure in HPC, LLVM-HPC 2022

作者： Tian, Shilei Huber, Joseph Parasyris, Konstantinos Chapman, Barbara Doerfert, Johannes Stony Brook University Department of Computer Science Stony Brook United States Advanced Micro Devices Austin United States Lawrence Livermore National Laboratory Center for Applied Scientific Computing Livermore United States

ISBN: (纸本)9781665475587

Currently, offloading to accelerators requires users to identify which regions are to be executed on the device, what memory needs to be transferred, and how synchronization is to be resolved. On top of these manual tasks, many standard (C/C++ library) functions, such as file I/O or memory manipulation, cannot be directly executed on the device and need to be worked around by the user explicitly. This makes it challenging to port programs in the first place and hinders developers from testing features on the GPU and within the GPU compilation pipeline. Existing tests and test suites for the host are effectively unusable for accelerators and need to be manually ported to provide the same benefits for the devices as they do on the host. In this paper, we propose a direct GPU compilation scheme that leverages the portable target offloading interface provided by LLVM/OpenMP. Utilizing this infrastructure allows us to compile an existing host application for the GPU and execute it there with only a minimal wrapper layer for the user code, command line arguments, and a compiler provided GPU implementation of C/C++ standard library functions. The C/C++ library functions are partially implemented for direct device execution and otherwise fallback to remote procedure call (RPC) to call host functions transparently. Our proposed prototype will allow users to quickly compile for, and test on, the GPU without explicitly handling kernel launches, data mapping, or host-device synchronization. We evaluate our implementation using three proxy applications with host OpenMP parallelism and three microbenchmarks to test the correctness of our prototype GPU compilation. © 2022 IEEE.

关键词： Graphics processing unit

来源：评论

学校读者我要写书评

暂无评论

Data Augmentation Techniques for Medical Image Segmentation - A Review

Data Augmentation Techniques for Medical Image Segmentation ...

引用

2024 International Conference on computer and Applications, ICCA 2024

作者： Kozah, Naney Dornaika, Fadi Charafeddine, Jinan El Jaam, Jihad University of the Basque Country UPV/EHU Department of Computer Science and Artificial Intelligence San Sebastian Spain University of the Basque Country UPV/EHU Ikerbasque Basque Foundation for Science San Sebastian Spain Léonard de Vinci Pôle Universitaire Research Center Paris La Défense92916 France School of Computing and Data Liverpool J. Mores University Liverpool United Kingdom

ISBN: (纸本)9798350367560

Segmentation of medical images is crucial for accu-rate diagnosis, treatment planning and intervention. However, the scarcity of data and the variability of different imaging modalities, including MRI, CT, PET and ultrasound, pose a major challenge for deep learning models. This review discusses traditional and advanced data augmentation techniques used to improve segmentation performance. In addition to geometric transformations such as rotation and scaling, advanced approaches such as Generative Adversarial Networks (GANs) and superpixel-based augmentations are also evaluated. The review also highlights novel methods such as Local-and-Contour Aware Grid Mixing (LCAMix) and Hard and Soft Mixing (HSMix) that preserve anatomical boundaries and effectively augment datasets. The analysis concludes with practical insights for clinical applications and suggests integrating these techniques into data pipelines to increase the generalizability of models. Future research directions focus on DICOM-specific augmentation strategies and optimization of computational requirements to improve clinical applicability. © 2024 IEEE.

关键词： Image segmentation

来源：评论

学校读者我要写书评

暂无评论

GeckoDAG: Towards a Lightweight DAG-Based Blockchain via Reducing Data Redundancy

GeckoDAG: Towards a Lightweight DAG-Based Blockchain via Red...

引用

International Conference on Distributed computing Systems

作者： Xiaohai Dai Yifan Zhou Jiang Xiao Feng Cheng Xia Xie Hai Jin Bo Li National Engineering Research Center for Big Data Technology and System Services Computing Technology and System Lab Cluster and Grid Computing Lab School of Computer Science and Technology Huazhong University of Science and Technology China School of Computer Science and Technology Hainan University China Department of Computer Science and Engineering Hong Kong University of Science and Technology Hong Kong

To overcome the scaling and performance limitations, the Directed Acyclic Graph (DAG) is utilized as the underlying storage model of blockchain systems, which enables concurrent transaction processing and confirmation. However, accompanied by high performance, DAG-based blockchains still suffer from the severe challenge of constrained storage scalability, i.e., expensive storage overhead. Based on an in-depth analysis of the data, we discover that the root cause of storage overhead stems from the considerable data redundancy in the DAG-based blockchains. In this paper, we propose GeckoDAG, a lightweight DAG-based blockchain, whose design consists of two steps. First, we abstract a storage model named Basic from the existing DAG-based blockchain systems, which offers both high performance and security. On top of Basic, we then devise GeckoDAG, which merges previous transactions into Transaction Union (TU) and reduces the data redundancy in TU, thus lowering the storage overhead. To evaluate our design, we implement a prototype of GeckoDAG and conduct various experiments on it. The experimental results demonstrate that GeckoDAG can offer storage scalability while maintaining the security and efficiency of DAG-based blockchains.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Coronary Artery Segmentation from X-ray Angiographic Images using Width-aware U-Net 2

Coronary Artery Segmentation from X-ray Angiographic Images ...

引用

2nd International Conference on computer Graphics, Image and Virtualization, ICCGIV 2022

作者： Song, Shuang Xiao, Ruoxiu Lu, Tong Zhu, Hang Beihang University School of Computer Science and Engineering Beijing China University of Science and Technology Beijing School of Computer and Communication Engineering Beijing China Visual 3D Medical Science and Technology Development Co. Ltd Beijing China The Sixty Medical Center of People's Liberation Army General Hospital Senior Department of Cardiology Beijing China

ISBN: (纸本)9781665492508

X-ray angiographic images are regarded as the gold standard in the intra-operation of vascular diseases due to their high imaging quality and speed. However, due to many vascular- similar structures existing in the images, the accurate segmentation of the coronary artery is challenging. In this paper, we proposed a width-aware U-Net model by supervising the segmentation procedure using the label mask and centerline- derived width map. In our model, the centerlines of label masks are extracted and width maps are then computed. Together with the label mask, the width maps are then utilized to supervise the segmentation network. By comparing the proposed network with other state-of-the-art models, we can achieve superior results in discarding non-vascular artefacts and preserving the integrity of the whole vascular tree. © 2022 IEEE.

关键词： Image segmentation

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：