检索结果-内蒙古大学图书馆

SegMAN: Omni-scale Context Modeling with State Space Models and Local Attention for Semantic Segmentation

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Fu, Yunxiang Lou, Meng Yu, Yizhou School of Computing and Data Science The University of Hong Kong Hong Kong

High-quality semantic segmentation relies on three key capabilities: global context modeling, local detail encoding, and multi-scale feature extraction. However, recent methods struggle to possess all these capabilities simultaneously. Hence, we aim to empower segmentation networks to simultaneously carry out efficient global context modeling, high-quality local detail encoding, and rich multi-scale feature representation for varying input resolutions. In this paper, we introduce SegMAN, a novel linear-time model comprising a hybrid feature encoder dubbed SegMAN Encoder, and a decoder based on state space models. Specifically, the SegMAN Encoder synergistically integrates sliding local attention with dynamic state space models, enabling highly efficient global context modeling while preserving fine-grained local details. Meanwhile, the MMSCopE module in our decoder enhances multi-scale context feature extraction and adaptively scales with the input resolution. We comprehensively evaluate SegMAN on three challenging datasets: ADE20K, Cityscapes, and COCO-Stuff. For instance, SegMAN-B achieves 52.6% mIoU on ADE20K, outperforming SegNeXt-L by 1.6% mIoU while reducing computational complexity by over 15% GFLOPs. On Cityscapes, SegMAN-B attains 83.8% mIoU, surpassing SegFormer-B3 by 2.1% mIoU with approximately half the GFLOPs. Similarly, SegMAN-B improves upon VWFormerB3 by 1.6% mIoU with lower GFLOPs on the COCO-Stuff dataset. Our code is available at https://***/yunxiangfu2001/SegMAN © 2024, CC BY.

关键词： Semantic Segmentation

ASE: PRACTICAL ACOUSTIC SPEED ESTIMATION BEYOND DOPPLER VIA SOUND DIFFUSION FIELD

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Lyu, Sheng Wu, Chenshu School of Computing and Data Science The University of Hong Kong Hong Kong

Passive human speed estimation plays a critical role in acoustic sensing. Despite extensive study, existing systems, however, suffer from various limitations: First, previous acoustic speed estimation exploits Doppler Frequency Shifts (DFS) created by moving targets and relies on microphone arrays, making them only capable of sensing the radial speed within a constrained distance. Second, the channel measurement rate proves inadequate to estimate high moving speeds. To overcome these issues, we present ASE, an accurate and robust Acoustic Speed Estimation system on a single commodity microphone. We model the sound propagation from a unique perspective of the acoustic diffusion field, and infer the speed from the acoustic spatial distribution, a completely different way of thinking about speed estimation beyond prior DFS-based approaches. We then propose a novel Orthogonal Time-Delayed Multiplexing (OTDM) scheme for acoustic channel estimation at a high rate that was previously infeasible, making it possible to estimate high speeds. We further develop novel techniques for motion detection and signal enhancement to deliver a robust and practical system. We implement and evaluate ASE through extensive real-world experiments. Our results show that ASE reliably tracks walking speed, independently of target location and direction, with a mean error of 0.13 m/s, a reduction of 2.5x from DFS, and a detection rate of 97.4% for large coverage, e.g., free walking in a 4m × 4m room. We believe ASE pushes acoustic speed estimation beyond the conventional DFS-based paradigm and will inspire exciting research in acoustic sensing. © 2024, CC BY.

关键词： Channel estimation

SparX: A Sparse Cross-Layer Connection Mechanism for Hierarchical Vision Mamba and Transformer Networks

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Lou, Meng Fu, Yunxiang Yu, Yizhou School of Computing and Data Science The University of Hong Kong Hong Kong

Due to the capability of dynamic state space models (SSMs) in capturing long-range dependencies with linear-time computational complexity, Mamba has shown notable performance in NLP tasks. This has inspired the rapid development of Mamba-based vision models, resulting in promising results in visual recognition tasks. However, such models are not capable of distilling features across layers through feature aggregation, interaction, and selection. Moreover, existing cross-layer feature aggregation methods designed for CNNs or ViTs are not practical in Mamba-based models due to high computational costs. Therefore, this paper aims to introduce an efficient cross-layer feature aggregation mechanism for vision backbone networks. Inspired by the Retinal Ganglion Cells (RGCs) in the human visual system, we propose a new sparse cross-layer connection mechanism termed SparX to effectively improve cross-layer feature interaction and reuse. Specifically, we build two different types of network layers: ganglion layers and normal layers. The former has higher connectivity and complexity, enabling multi-layer feature aggregation and interaction in an input-dependent manner. In contrast, the latter has lower connectivity and complexity. By interleaving these two types of layers, we design a new family of vision backbone networks with sparsely cross-connected layers, achieving an excellent trade-off among model size, computational cost, memory cost, and accuracy in comparison to its counterparts. For instance, with fewer parameters, SparX-Mamba-T improves the top-1 accuracy of VMamba-T from 82.5% to 83.5%, while SparX-Swin-T achieves a 1.3% increase in top-1 accuracy compared to Swin-T. Extensive experimental results demonstrate that our new connection mechanism possesses both superior performance and generalization capabilities on various vision tasks. Code is publicly available at https://***/LMMMEng/SparX. © 2024, CC BY.

关键词： State space methods

Intelligent Refrigerator using Machine Learning and IoT

学校读者我要写书评

暂无评论

Intelligent Refrigerator using Machine Learning and IoT

2022 IEEE International Conference on Advances in computing, Communication and Applied Informatics, ACCAI 2022

作者： Rajeswari, D. Srinivasan, R. Ramamoorthy, S. Pushpalatha, M. School of Computing College of Engineering and Technology Department of Data Science and Business Systems India School of Computing College of Engineering and Technology SRM Institute of Science and Technology Department of Computing Technologies Kattankulathur603203 India

ISBN: (纸本)9781665495295

Machine Learning and Internet of Things (IoT) play a vital role in the current electronics world. Electronic appliances operated by humans are automated and accessed remotely via IoT and 5G networks. Smart Cars, Drone Deliveries, and Intelligent Refrigerators are some of IoT and machine learningmajor applications. Potentially, with the aid of computing devices Intelligent Refrigerators can be accessed from anywhere and anytime. In Intelligent Refrigerators,a single tap on the computing devices predicts the quality of food items kept in the refrigerator. Besides, in order to work smartly, intelligent IoT devices deployed in intelligent refrigerators predict food quality anonymously using machine learningtechniques. This paper uses ensemble machine learning technique with naive bayes, support vector machine and decision tree to classify the considered fruits are spoiled or not. © 2022 IEEE.

关键词： Support vector machines

Improved Particle Approximation Error for Mean Field Neural Networks

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Nitanda, Atsushi Singapore College of Computing and Data Science Nanyang Technological University Singapore

Mean-field Langevin dynamics (MFLD) minimizes an entropy-regularized nonlinear convex functional defined over the space of probability distributions. MFLD has gained attention due to its connection with noisy gradient descent for mean-field two-layer neural networks. Unlike standard Langevin dynamics, the nonlinearity of the objective functional induces particle interactions, necessitating multiple particles to approximate the dynamics in a finite-particle setting. Recent works (Chen et al., 2022;Suzuki et al., 2023b) have demonstrated the uniform-in-time propagation of chaos for MFLD, showing that the gap between the particle system and its mean-field limit uniformly shrinks over time as the number of particles increases. In this work, we improve the dependence on logarithmic Sobolev inequality (LSI) constants in their particle approximation errors which can exponentially deteriorate with the regularization coefficient. Specifically, we establish an LSI-constant-free particle approximation error concerning the objective gap by leveraging the problem structure in risk minimization. As the application, we demonstrate improved convergence of MFLD, sampling guarantee for the mean-field stationary distribution, and uniform-in-time Wasserstein propagation of chaos in terms of particle complexity. Copyright © 2024, The Authors. All rights reserved.

关键词： Multilayer neural networks

AnywhereDoor: Multi-Target Backdoor Attacks on Object Detection

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Lu, Jialin Shan, Junjie Zhao, Ziqi Chow, Ka-Ho School of Computing and Data Science The University of Hong Kong Hong Kong

As object detection becomes integral to many safety-critical applications, understanding its vulnerabilities is essential. Backdoor attacks, in particular, pose a serious threat by implanting hidden triggers in victim models, which adversaries can later exploit to induce malicious behaviors during inference. However, current understanding is limited to single-target attacks, where adversaries must define a fixed malicious behavior (target) before training, making inference-time adaptability impossible. Given the large output space of object detection (including object existence prediction, bounding box estimation, and classification), the feasibility of flexible, inference-time model control remains unexplored. This paper introduces AnywhereDoor, a multi-target backdoor attack for object detection. Once implanted, AnywhereDoor allows adversaries to make objects disappear, fabricate new ones, or mislabel them, either across all object classes or specific ones, offering an unprecedented degree of control. This flexibility is enabled by three key innovations: (i) objective disentanglement to scale the number of supported targets;(ii) trigger mosaicking to ensure robustness even against region-based detectors;and (iii) strategic batching to address object-level data imbalances that hinder manipulation. Extensive experiments demonstrate that AnywhereDoor grants attackers a high degree of control, improving attack success rates by 26% compared to adaptations of existing methods for such flexible control. Copyright © 2024, The Authors. All rights reserved.

关键词： Object detection

A Person-job Matching Method Based on BM25 and Pre-Trained Language Model 6

学校读者我要写书评

暂无评论

A Person-job Matching Method Based on BM25 and Pre-Trained L...

6th International Conference on Machine Learning and Natural Language Processing, MLNLP 2023

作者： Tang, Jiaqi Chen, Heyun Chen, Zifan Pan, Jiahua He, Yijiang Zhao, Jie Li, Zhen PKU-Changsha Institute for Computing and Digital Economy Changsha China Center for Data Science Peking University Beijing China National Engineering Laboratory for Big Data Analysis and Applications Peking University Beijing China

ISBN: (纸本)9798400709241

To address the inefficiency of traditional methods for person-job matching and the lack of interpretability of deep learning approaches, a novel approach for person-job matching based on BM25 and pre-Trained language models is proposed. First, the BM25 algorithm is utilized to perform an initial screening of resumes based on job descriptions. By calculating the matching degree of keywords, it rapidly filters out candidate resumes that do not align with the job requirements. Subsequently, the pre-Trained language model BERT is employed to perform fine-grained sorting of the resumes that passed the initial screening, identifying the candidates that best match with the job. By combining the screening capability of the BM25 algorithm with the semantic understanding ability of pre-Trained language models, this method enhances the accuracy and interpretability of person-job matching results while maintaining matching speed. Furthermore, to improve model generalization, large-scale language models are employed for data augmentation. Experimental results indicate the significant potential for the proposed approach in practical applications, offering improved solutions for the challenges of talent recruitment across diverse job demands. © 2023 ACM.

关键词： Semantics

Prompt Engineering for Curriculum Design

学校读者我要写书评

暂无评论

Prompt Engineering for Curriculum Design

Educational Technology (ICET), IEEE International Conference on

作者： Jonathan Leung Zhiqi Shen College of Computing and Data Science Nanyang Technological University Singapore Singapore

ISBN: (数字)9798350376944

ISBN: (纸本)9798350376951

Large Language Models (LLMs) have recently demonstrated successes in a broad range of areas that require language understanding and generation. The design of course curricula is a key part of the educational process, as it helps students understand expectations and learning goals, and teachers to maintain consistency when conducting the course. Thus, there would be benefits if educators could leverage the capabilities of LLMs. The first benefit is the potential for improving course content, which would lead to better learning outcomes for students. Secondly, educators can save time by having an LLM assist in creating course material or assessments. However, many methods of using LLMs in specific applications fine-tune LLMs, which requires task-specific data as well as technical knowledge to perform the fine-tuning. In this work we use Prompt Engineering, a set of methods that aims to improve generated responses from LLMs by altering the input prompt to the LLM. Often, however, advice for designing prompts is very broad, such as "be concise". We utilize prompt patterns to create and propose three reusable, specific, and customizable prompts that can be used to assist in the design of a course syllabus, lesson material, and assessment questions. The use of prompt patterns additionally aims to produce reliable and consistent results from LLMs.

关键词： Knowledge engineering Visualization Image synthesis Large language models Educational technology Chatbots Prompt engineering Reliability

Intelligent Blocking and Prevention of SARS-CoV-2 Based on Evolutionary Reinforcement Learning 13

学校读者我要写书评

暂无评论

Intelligent Blocking and Prevention of SARS-CoV-2 Based on E...

13th IEEE Congress on Evolutionary Computation, CEC 2024

作者： He, Chengwei Hong, Huaping Ma, Lijia Zhu, Qingling Bai, Yuan College of Computer Science and Software Engineering Shenzhen University Shenzhen518060 China Shenzhen University National Engineering Laboratory for Big Data System Computing Technology Shenzhen518060 China

ISBN: (纸本)9798350308365

In recent years, the global repercussions of SARS-CoV-2 and its variants have posed significant challenges to various areas, including the economic order, transportation, healthcare, and education, and the mitigation and prevention of SARS-CoV-2 and other infectious diseases have received much attention. This paper tries to develop a rational and effective intervention strategy in response to the dynamic evolution of the SARS-CoV-2 epidemic. To address this issue, we first use the SEIR model to analyze the population distribution under the SARS-CoV-2 epidemic, and then integrate a deep Q-neural network (DQN) to reflect an economic quantification model contextualized within the SARS-CoV-2 scenario. Next, we propose an evolutionary deep reinforcement learning to generate an optimal and adaptive intervention strategy, by optimizing the weight sequence of the DQN. The experimental results show the effectiveness of the proposed method in maximizing economic benefits, while ensuring the uninterrupted functionality of medical institutions. Importantly, the learned intervention strategy exhibits the applicability to other infectious diseases. © 2024 IEEE.

关键词： Deep reinforcement learning