检索结果-内蒙古大学图书馆

12th International Conference on Learning Representations, ICLR 2024

作者： Kou, Siqi Gan, Lei Wang, Dequan Li, Chongxuan Deng, Zhijie Qing Yuan Research Institute SEIEE Shanghai Jiao Tong University China Gaoling School of Artificial Intelligence Renmin University of China China Beijing Key Laboratory of Big Data Management and Analysis Methods China School of Computer Science Fudan University China Shanghai Artificial Intelligence Laboratory China

Diffusion models have impressive image generation capability, but low-quality generations still exist, and their identification remains challenging due to the lack of a proper sample-wise metric. To address this, we propose BayesDiff, a pixel-wise uncertainty estimator for generations from diffusion models based on Bayesian inference. In particular, we derive a novel uncertainty iteration principle to characterize the uncertainty dynamics in diffusion, and leverage the last-layer Laplace approximation for efficient Bayesian inference. The estimated pixel-wise uncertainty can not only be aggregated into a sample-wise metric to filter out low-fidelity images but also aids in augmenting successful generations and rectifying artifacts in failed generations in text-to-image tasks. Extensive experiments demonstrate the efficacy of BayesDiff and its promise for practical applications. Our code is available at https://***/karrykkk/BayesDiff. © 2024 12th International Conference on Learning Representations, ICLR 2024. All rights reserved.

关键词： Pixels

来源：评论

学校读者我要写书评

暂无评论

MEMORY-ASSISTED SUB-PROTOTYPE MINING FOR UNIVERSAL DOMAIN ADAPTATION 12

MEMORY-ASSISTED SUB-PROTOTYPE MINING FOR UNIVERSAL DOMAIN AD...

引用

12th International Conference on Learning Representations, ICLR 2024

作者： Lai, Yuxiang Zhou, Yi Liu, Xinghong Zhou, Tao School of Computer Science and Engineering Southeast University China Key Laboratory of New Generation Artificial Intelligence Technology and Its Interdisciplinary Applications Southeast University Ministry of Education China School of Computer Science and Engineering Nanjing University of Science and Technology China

Universal domain adaptation aims to align the classes and reduce the feature gap between the same category of the source and target domains. The target private category is set as the unknown class during the adaptation process, as it is not included in the source domain. However, most existing methods overlook the intra-class structure within a category, especially in cases where there exists significant concept shift between the samples belonging to the same category. When samples with large concept shifts are forced to be pushed together, it may negatively affect the adaptation performance. Moreover, from the interpretability aspect, it is unreasonable to align visual features with significant differences, such as fighter jets and civil aircraft, into the same category. Unfortunately, due to such semantic ambiguity and annotation cost, categories are not always classified in detail, making it difficult for the model to perform precise adaptation. To address these issues, we propose a novel Memory-Assisted Sub-Prototype Mining (MemSPM) method that can learn the differences between samples belonging to the same category and mine sub-classes when there exists significant concept shift between them. By doing so, our model learns a more reasonable feature space that enhances the transferability and reflects the inherent differences among samples annotated as the same category. We evaluate the effectiveness of our MemSPM method over multiple scenarios, including UniDA, OSDA, and PDA. Our method achieves state-of-the-art performance on four benchmarks in most cases. © 2024 12th International Conference on Learning Representations, ICLR 2024. All rights reserved.

关键词： Benchmarking

来源：评论

学校读者我要写书评

暂无评论

Matching Gains with Pays: Effective and Fair Learning in Multi-Agent Public Goods Dilemmas 27

Matching Gains with Pays: Effective and Fair Learning in Mul...

引用

27th European Conference on artificial intelligence, ECAI 2024

作者： Chen, Yitian Liu, Xuan Zhang, Shigeng Chen, Xinning Guo, Song School of Computer Science and Engineering Central South University China College of Computer Science and Electronic Engineering Hunan University China The Ministry of Education Key Laboratory of "Fusion Computing of Supercomputing and Artificial Intelligence" China Department of Computer Science and Engineering Hong Kong University of Science and Technology China

ISBN: (纸本)9781643685489

The training of multi-agent reinforcement learning (MARL) tasks with the public goods dilemma (PGD) is difficult because the selfish actions of individual agents for high personal rewards may reduce the collective utility of the whole group. Existing solutions to this problem, e.g., reward gifting or intrinsic rewards, although inducing cooperation among agents in small groups, cannot guarantee fairness among agents’ policies and fail to achieve optimal group utility in large-scale systems. In this paper, we propose F4PGD, an effective method to train large-scale MARL tasks with PGD in a decentralized manner, which is inspired by Adam’s equity theory that the match between a person’s payoff and his contribution is the key incentive for people to contribute to the common good. In F4PGD, a mechanism is designed to match an agent’s reward with its contribution, which suppresses agents from taking a free ride and meanwhile encourages well-learned agents to contribute to public goods. Experimental results show that F4PGD effectively learns optimal policies for the whole group and guarantees fairness among agents in several typical MARL tasks with PGD. © 2024 The Authors.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Text-Guided Attention is All You Need for Zero-Shot Robustness in Vision-Language Models 38

Text-Guided Attention is All You Need for Zero-Shot Robustne...

引用

38th Conference on Neural Information Processing Systems, NeurIPS 2024

作者： Yu, Lu Zhang, Haiyang Xu, Changsheng School of Computer Science and Engineering Tianjin University of Technology China State Key Laboratory of Multimodal Artificial Intelligence Systems Institute of Automation University of Chinese Academy of Sciences China

Due to the impressive zero-shot capabilities, pre-trained vision-language models (e.g. CLIP), have attracted widespread attention and adoption across various domains. Nonetheless, CLIP has been observed to be susceptible to adversarial examples. Through experimental analysis, we have observed a phenomenon wherein adversarial perturbations induce shifts in text-guided attention. Building upon this observation, we propose a simple yet effective strategy: Text-Guided Attention for Zero-Shot Robustness (TGA-ZSR). This framework incorporates two components: the Attention Refinement module and the Attention-based Model Constraint module. Our goal is to maintain the generalization of the CLIP model and enhance its adversarial robustness: The Attention Refinement module aligns the text-guided attention obtained from the target model via adversarial examples with the text-guided attention acquired from the original model via clean examples. This alignment enhances the model's robustness. Additionally, the Attention-based Model Constraint module acquires text-guided attention from both the target and original models using clean examples. Its objective is to maintain model performance on clean samples while enhancing overall robustness. The experiments validate that our method yields a 9.58% enhancement in zero-shot robust accuracy over the current state-of-the-art techniques across 16 datasets. Our code is available at https://***/zhyblue424/TGA-ZSR. © 2024 Neural information processing systems foundation. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Reassessing Glove Embeddings in Deep Learning: A Comparative Study with Classical ML Approaches 15

Reassessing Glove Embeddings in Deep Learning: A Comparative...

引用

15th International Conference on Emerging Ubiquitous Systems and Pervasive Networks / 14th International Conference on Current and Future Trends of Information and Communication Technologies in Healthcare, EUSPN/ICTH 2024

作者： Labd, Zakia Bahassine, Said Housni, Khalid Aadi, Fatima Zahrae Aithamou Laboratory of research in informatics LatRI Department of Computer Science Faculty of Sciences Ibn Tofail University Kenitra Morocco Laboratory of Artificial Intelligence and Complex Systems Engineering Department of Computer Science National Higher School of Arts and Crafts Hassan II University Casablanca Morocco

The importance of text classification algorithms has increased due to the growing availability of large-scale data. This has led to a greater demand for efficient classification techniques and encoding algorithms. Word embedding techniques, like Glove, have shown significant success in encoding semantic relationships between words. This research paper aims to reassess the effectiveness of Glove embeddings coupled with deep learning algorithms. The impact of Glove embedding on two widely used deep learning models: Recurrent Neural Networks (RNN) and Recurrent Convolutional Neural Networks (RCNN) is analyzed. The results highlight the impact of Glove embeddings on deep learning models, showcasing significant performance enhancements in some cases while having minimal effects in others. By examining the impact of Glove embeddings on traditional ML algorithms in a previous study, valuable context for understanding the performance differences between the two approaches is obtained. © 2024 The Authors. Published by Elsevier B.V.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

An Indicator Based Evolutionary Algorithm for Multiparty Multiobjective Knapsack Problems 13th

An Indicator Based Evolutionary Algorithm for Multiparty Mu...

引用

13th IFIP TC 12 International Conference on Intelligent Information Processing, IIP 2024

作者： Song, Zhen Luo, Wenjian Xu, Peilan Ye, Zipeng Chen, Kesheng Guangdong Provincial Key Laboratory of Novel Security Intelligence Technologies School of Computer Science and Technology Harbin Institute of Technology Shenzhen518055 China Peng Cheng Laboratory Shenzhen518000 China School of Artificial Intelligence Nanjing University of Information Science and Technology Nanjing210044 China

ISBN: (纸本)9783031578076

As a special case of the multiobjective optimization problem, the multiobjective knapsack problem (MOKP) widely exists in real-world applications. Currently, most algorithms used to solve MOKPs assume that these problems involve only one decision maker (DM). However, some complex MOKPs often involve more than one decision makers and we call such problems multiparty multiobjective knapsack problems (MPMOKPs). Existing algorithms cannot solve MPMOKPs effectively. To the best of our knowledge, there is only a little attention paid to MPMOKPs. In this paper, inspired by existing SMS-EMOA, we propose a novel indicator-based algorithm called SMS-MPEMOA to solve MPMOKPs, which aims to search solutions to satisfy all decision makers as much as possible. SMS-MPEMOA is compared with several state-of-the-art multiparty multiobjective optimization algorithms (MPMOEAs) on the benchmarks and the experimental results demonstrate that SMS-MPEMOA is very competitive. © IFIP International Federation for Information Processing 2024.

关键词： Multiobjective optimization

来源：评论

学校读者我要写书评

暂无评论

Object Detection for Retail Product Recognition 9

Object Detection for Retail Product Recognition

引用

9th International Conference on Business and Industrial Research, ICBIR 2024

作者： Pannoy, Nakul Nonsiri, Sarayut Makdee, Supawee Thai-Nichi Institute of Technology Artificial Intelligence and Internet of Things Research Laboratory Bangkok Thailand Ubon Ratchathani Rajabhat University Faculty of Computer Science Ubon Ratchathani Thailand

ISBN: (纸本)9798350383010

The retail sector is a vital driver of economic growth. The retail industry must adopt technology to enhance productivity, streamline operations, and minimize human errors in order to continue its crucial economic role. During the COVID-19 pandemic, the purchasing power volumes globally grew from February 2020 to April 2021, and the retail sector gained 35 percents in market capitalization. This data highlights a compelling research opportunity. Consequently, artificial intelligence (AI) has become a focus of significant interest and has widely adopted technology, which includes computer vision to recognize and detect retail products. This paper aims to explore YOLO's performance in retail product recognition. The research compares different YOLO versions to identify on-shelf grocery items. In this research, YOLOv8 is applied and divided into five subcategories: YOLOv8n (nano), YOLOv8s (small), YOLOv8m (medium), YOLOv81 (large), and YOLOv8x (extra-large). They were evaluated with Grozi-120 and SKUII0K datasets. The evaluation metrics are precision, recall, mAP50, and mAP50-95. The result shows that YOLOv8x provides the best overall performance in both datasets, where mAP50 metrics exhibits the highest score at 92.6 percents in SKUII0K. The outcomes show that the YOLOv8 model works well for retail product detection. © 2024 IEEE.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

SAFE OFFLINE REINFORCEMENT LEARNING WITH FEASIBILITY-GUIDED DIFFUSION MODEL 12

SAFE OFFLINE REINFORCEMENT LEARNING WITH FEASIBILITY-GUIDED ...

引用

12th International Conference on Learning Representations, ICLR 2024

作者： Zheng, Yinan Li, Jianxiong Yu, Dongjie Yang, Yujie Li, Shengbo Eben Zhan, Xianyuan Liu, Jingjing Tsinghua University China School of Vehicle and Mobility Tsinghua University China Department of Computer Science The University of Hong Kong Hong Kong Shanghai Artificial Intelligence Laboratory China

Safe offline reinforcement learning is a promising way to bypass risky online interactions towards safe policy learning. Most existing methods only enforce soft constraints, i.e., constraining safety violations in expectation below thresholds predetermined. This can lead to potentially unsafe outcomes, thus unacceptable in safety-critical scenarios. An alternative is to enforce the hard constraint of zero violation. However, this can be challenging in offline setting, as it needs to strike the right balance among three highly intricate and correlated aspects: safety constraint satisfaction, reward maximization, and behavior regularization imposed by offline datasets. Interestingly, we discover that via reachability analysis of safe-control theory, the hard safety constraint can be equivalently translated to identifying the largest feasible region given the offline dataset. This seamlessly converts the original trilogy problem to a feasibility-dependent objective, i.e., maximizing reward value within the feasible region while minimizing safety risks in the infeasible region. Inspired by these, we propose FISOR (FeasIbility-guided Safe Offline RL), which allows safety constraint adherence, reward maximization, and offline policy learning to be realized via three decoupled processes, while offering strong safety performance and stability. In FISOR, the optimal policy for the translated optimization problem can be derived in a special form of weighted behavior cloning, which can be effectively extracted with a guided diffusion model thanks to its expressiveness. Moreover, we propose a novel energy-guided sampling method that does not require training a complicated time-dependent classifier to simplify the training. We compare FISOR against baselines on DSRL benchmark for safe offline RL. Evaluation results show that FISOR is the only method that can guarantee safety satisfaction in all tasks, while achieving top returns in most tasks. Project website: https://zhengyinan

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

A Civil Aviation Customer Service Ontology and Its Applications

引用

Data intelligence 2023年第4期5卷 1063-1081页

作者： Meixiang Lv Xudong Cao Tianxing Wu Yuehua Li School of Computer Science and Engineering Southeast UniversityNanjingChina Key Laboratory of New Generation Artificial Intelligence Technology and Its Interdisciplinary Applications(Southeast University) Ministry of EducationChina Zhejiang Lab HangzhouChina

In the process of developing the C919 large aircraft customer service intelligence system,we find that heterogeneous and incomplete data cause the inefficient and inaccurate decision ***,to solve this problem,we propose to introduce the idea of ontology modeling and reasoning into competitive intelligence system building in this *** first present the building principles and methods of the civil aviation customer service *** then define the classes and properties to contribute a real-world civil aviation customer service ontology,which is published on the Web(http:/***/dataset/cacso).We finally design SWRL rules corresponding to different intelligence analysis targets to support reasoning in our designed competitive intelligence system.

关键词： Ontology Building intelligence Service Civil Aviation Customer Service

来源：评论

学校读者我要写书评

暂无评论

Local-Global Features Fusion Network for Distinguishing Radiolucent Jaw Lesions via CBCT

Local-Global Features Fusion Network for Distinguishing Radi...

引用

2024 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2024

作者： Wang, Yuan Chen, Hua Cai, Zikang Mao, Liang Sun, Zhijun Liu, Juan Wuhan University Institute of Artificial Intelligence School of Computer Science Wuhan China Wuhan University Key Laboratory of Oral Biomedicine Moe School and Hospital of Stomatology Wuhan China

ISBN: (纸本)9798350386226

Accurately distinguishing different types of jaw-bone radiation lesions (RJLs) based on cone beam computed tomography (CBCT) images is crucial for oral surgeons to choose appropriate treatment plans. Currently, only experienced radiologists can distinguish different types of RJLs from CBCT images. Therefore, it is necessary to study computing methods for accurately classifying CBCT images of different JRLs lesions. However, the lesions in CBCT images are very small, and different types of lesions exhibit high similarity on the images, posing challenges for computing methods. In this paper, we propose a Local-Global Features Fusion Network (LGFFNet) that simultaneously extracts local features and global features related to the lesions in CBCT images and fuses them. In order to distinguish different types of high similarity CBCT images, we adopted a loss function by combing focal loss and cross entropy loss to train the model, so that the model focuses more on learning the features of difficult-to-classify images while learning general distinguishing features. The experimental results on our collected dataset show that the classification performance of our method is superior to other comparative methods. © 2024 IEEE.

关键词： computerized tomography

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：