检索结果-内蒙古大学图书馆

Unified-modal Salient Object Detection via Adaptive Prompt Learning

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Wang, Kunpeng Li, Chenglong Tu, Zhengzheng Liu, Zhengyi Luo, Bin Information Materials and Intelligent Sensing Laboratory of Anhui Province Anhui Provincial Key Laboratory of Multimodal Cognitive Computation School of Computer Science and Technology Anhui University Hefei230601 China Key Laboratory of Intelligent Computing and Signal Processing of Ministry of Education School of Computer Science and Technology Anhui University Hefei China Anhui Provincial Key Laboratory of Multimodal Cognitive Computation School of Artificial Intelligence Anhui University Hefei230601 China The Institute of Physical Science and Information Technology Anhui University Hefei230601 China

Existing single-modal and multi-modal salient object detection (SOD) methods focus on designing specific architectures tailored for their respective tasks. However, developing completely different models for different tasks leads to labor and time consumption, as well as high computational and practical deployment costs. In this paper, we attempt to address both single-modal and multi-modal SOD in a unified framework called UniSOD, which fully exploits the overlapping prior knowledge between different tasks. Nevertheless, assigning appropriate strategies to modality variable inputs is challenging. To this end, UniSOD learns modality-aware prompts with task-specific hints through adaptive prompt learning, which are plugged into the proposed pre-trained baseline SOD model to handle corresponding tasks, while only requiring few learnable parameters compared to training the entire model. Each modality-aware prompt is generated from a switchable prompt generation block, which adaptively performs structural switching based on single-modal and multi-modal inputs without human intervention. Through end-to-end joint training, UniSOD achieves overall performance improvement on 14 benchmark datasets for RGB, RGB-D, and RGB-T SOD, which demonstrates that our method effectively and efficiently unifies single-modal and multi-modal SOD tasks. The code and results are available at https://***/Angknpng/UniSOD. Copyright © 2023, The Authors. All rights reserved.

关键词： Object detection

A Better Variant of self-attention for image caption

学校读者我要写书评

暂无评论

SSRN

SSRN 2022年

作者： Chen, Haorong Shi, YueXiang Guo, Lijun Key Laboratory of Intelligent Computing and Information Processing Ministry of Education Xiangtan University China School of Computer and Cyberspace Security Xiangtan University China Hunan Ordnance and Light Weapons Research Institute Co. Ltd China

The image description task aims to describe the content contained in the image using natural language. In the research field of computer vision and natural language processing, the image description task has attracted extensive attention from the academic community, especially after adding the attention mechanism, especially the self attention, which plays an increasingly important role in vision and natural language processing, Self attention updates the features of each location by calculating the weighted sum of features using the paired affinity of all locations. However, this ignores the potential correlation between different samples. In this paper, a new attention mechanism is proposed. Query and key are traditional inputs, while its Value is an external, small, and learnable shared memory, which can be realized by using the linkage layer and the normalization layer;It conveniently replaces the self attention in the existing popular architecture. External concerns have linear complexity and implicitly consider the correlation between all data samples. In this paper, we add our method to image capture to improve the accuracy of the image description model based on attention fusion. The model was evaluated on MSCOCO and Flickr30k datasets. The Cider score of MSCOCO datasets reached 130.7, which was superior to similar image description generation models. © 2022, The Authors. All rights reserved.

关键词： Reinforcement learning

Distributed Traffic Engineering in Hybrid Software Defined Networks: A Multi-agent Reinforcement Learning Framework

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Guo, Yingya Tang, Qi Ma, Yulong Tian, Han Chen, Kai The College of Computer and Data Science Fuzhou University China The Fujian Provincial Key Laboratory of Network Computing and Intelligent Information Processing Fuzhou University China The Key Laboratory of Spatial Data Mining & Information Sharing Ministry of Education Fujian 350003 China The Department of Computer Science & Engineering Hong Kong University of Science & Technology Hong Kong

Traffic Engineering (TE) is an efficient technique to balance network flows and thus improves the performance of a hybrid Software Defined Network (SDN). Previous TE solutions mainly leverage heuristic algorithms to centrally optimize link weight setting or traffic splitting ratios under the static traffic demand. Note that as the network scale becomes larger and network management gains more complexity, it is notably that the centralized TE methods suffer from a high computation overhead and a long reaction time to optimize routing of flows when the network traffic demand dynamically fluctuates or network failures happen. To enable adaptive and efficient routing in TE, we propose a Multi-agent Reinforcement Learning method CMRL that divides the routing optimization of a large network into multiple small-scale routing decision-making problems. To coordinate the multiple agents for achieving a global optimization goal, we construct an interactive environment for training the routing agents that own partial link utilization observations. To optimize credit assignment of multi-agent, we introduce the difference reward assignment mechanism for encouraging agents to take better action. Extensive simulations conducted on the real traffic traces demonstrate the superiority of CMRL in improving TE performance, especially when traffic demands change or network failures happen. Copyright © 2023, The Authors. All rights reserved.

关键词： Global optimization

An effective method using duty cycles for assessment of exposure to mobile communication systems

学校读者我要写书评

暂无评论

IEEE Electromagnetic Compatibility Magazine 2019年第2期8卷 57-63页

作者： Yang, Wanchun Zhou, Kai Tan, Pingan Gao, Xieping College of Information Engineering Key Laboratory of Intelligent Computing Information Processing of Education Ministry Xiangtan University China

An effective method for assessment of exposure to the electric field of the wireless signals from mobile communication systems is still lack, since the transmission of wireless signals is related to the traffic and the exposure level varies with the traffic load. In this paper we study the allocation mechanism of the traffic time slots in the mobile communication systems, and analyze the probability of the number of simultaneously occupied traffic time slots. Based on the queuing theory, the duty cycle of the wireless signal can be calculated by the traffic load. We establish the theoretical relationship between the exposure level and the traffic load. The exposure level in a certain period can be directly extrapolated by the calculated duty cycle, and the maximum electric field measured in the max-hold mode. By this method the maximum electric field only needs to be measured one time. Much effort on measurement can be avoided. © 2012 IEEE.

关键词： Electromagnetic fields

A Weighted Flat Lattice Transformer-based Knowledge Extraction Architecture for Chinese Named Entity Recognition

学校读者我要写书评

暂无评论

A Weighted Flat Lattice Transformer-based Knowledge Extracti...

International Conference on Computer Supported Cooperative Work in Design

作者： Hengwei Zhang Yuejia Wu Jian-Tao Zhou Inner Mongolia University College of Computer Science Hohhot China Ministry of Education National & Local Joint Engineering Research Center of Intelligent Information Processing Technology for Mongolian Engineering Research Center of Ecological Big Data Hohhot China Inner Mongolia Engineering Laboratory for Cloud Computing and Service Software Hohhot China Inner Mongolia Key Laboratory of Social Computing and Data Processing Hohhot China Inner Mongolia Key Laboratory of Discipline Inspection and Supervision Big Data Hohhot China Inner Mongolia Engineering Laboratory for Big Data Analysis Technology Hohhot China

ISBN: (数字)9798350349184

ISBN: (纸本)9798350349191

Named Entity Recognition (NER) is one of the contents of Knowledge Extraction (KE) that transforms data into knowledge representation. However, Chinese NER faces the problem of lacking clear word boundaries that limit the effectiveness of the KE. Although the flat lattice Transformer (FLAT) framework, which converts lattice structure into a flat structure including a set of spans, can effectively improve this problem and obtain advanced results, there still exist the problems of insensitivity to entity importance weights and insufficient feature learning. This paper proposes a weighted flat lattice Transformer architecture for Chinese NER, namely WFLAT. The WFLAT first adds a weight matrix into self-attention calculation, which can obtain finer-grained partitioning of entities to improve experimental performance, and then adopts a multi-layer Transformer encoder with each layer using a multi-head self-attention mechanism. Extensive experimental results on benchmarks demonstrate that our proposed KE model can obtain state-of-the-art performance for the Chinese NER task.

关键词： Representation learning Computational modeling Memory management Lattices Named entity recognition Transforms Knowledge representation

Quantum simulation of the Fokker-Planck equation via Schrödingerization

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Jin, Shi Liu, Nana Yu, Yue School of Mathematical Sciences Institute of Natural Sciences MOE-LSC Shanghai Jiao Tong University Shanghai200240 China University of Michigan Shanghai Jiao Tong University Joint Institute Shanghai200240 China School of Mathematics and Computational Science Hunan Key Laboratory for Computation and Simulation in Science and Engineering Key Laboratory of Intelligent Computing and Information Processing of Ministry of Education Xiangtan University Hunan Xiangtan411105 China

This paper studies a quantum simulation technique for solving the Fokker-Planck equation. Traditional semi-discretization methods often fail to preserve the underlying Hamiltonian dynamics and may even modify the Hamiltonian structure, particularly when incorporating boundary conditions. We address this challenge by employing the Schrödingerization method - it converts any linear partial and ordinary differential equation with non-Hermitian dynamics into systems of Schrödinger-type equations. It does so via the so-called warped phase transformation that maps the equation into one higher dimension. We explore the application in two distinct forms of the Fokker-Planck equation. For the conservation form, we show that the semi-discretization-based Schrödingerization is preferable, especially when dealing with non-periodic boundary conditions. Additionally, we analyze the Schrödingerization approach for unstable systems that possess positive eigenvalues in the real part of the coefficient matrix or differential operator. Our analysis reveals that the direct use of Schrödingerization has the same effect as a stabilization procedure. For the heat equation form, we propose a quantum simulation procedure based on the time-splitting technique, and give explicitly its corresponding quantum circuit. We discuss the relationship between operator splitting in the Schrödingerization method and its application directly to the original problem, illustrating how the Schrödingerization method accurately reproduces the time-splitting solutions at each step. Furthermore, we explore finite difference discretizations of the heat equation form using shift operators. Utilizing Fourier bases, we diagonalize the shift operators, enabling efficient simulation in the frequency space. Providing additional guidance on implementing the diagonal unitary operators, we conduct a comparative analysis between diagonalizations in the Bell and the Fourier bases, and show that the former generally exhibit

关键词： Hamiltonians

Quasi-cyclic perfect codes in Doob graphs and special partitions of Galois rings

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Shi, Minjia Li, Xiaoxiao Krotov, Denis S. Özbudak, Ferruh Key Laboratory of Intelligent Computing and Signal Processing Ministry of Education School of Mathematical Sciences Anhui University Hefei230601 China School of Mathematical Sciences Anhui University Hefei230601 China Sobolev Institute of Mathematics pr. Akademika Koptyuga 4 Novosibirsk630090 Russia Faculty of Engineering and Natural Sciences Sabancı University Middle East Technical University 34956 Istanbul Ankara06800 Turkey

The Galois ring GR(4∆) is the residue ring Z4[x]/(h(x)), where h(x) is a basic primitive polynomial of degree ∆ over Z4. For any odd ∆ larger than 1, we construct a partition of GR(4∆)\{0} into 6-subsets of type {a, b, −a−b, −a, −b, a+b} and 3-subsets of type {c, −c, 2c} such that the partition is invariant under the multiplication by a nonzero element of the Teichmuller set in GR(4∆) and, if ∆ is not a multiple of 3, under the action of the automorphism group of GR(4∆). As a corollary, this implies the existence of quasi-cyclic additive 1-perfect codes of index (2∆ − 1) in D((2∆ − 1)(2∆ − 2)/6, 2∆ − 1) where D(m, n) is the Doob metric scheme on Z2m+n .MSC Codes 94B99 Copyright © 2023, The Authors. All rights reserved.

关键词： Algebra

Component edge connectivity of hypercube-like networks

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Liu, Dong Li, Pingshan Zhang, Bicheng Key Laboratory of Intelligent Computing & Information Processing of Ministry of Education School of Mathematical and Computational Sicence Xiangtan University Hunan Xiangtan41105 China

As a generalization of the traditional connectivity, the g-component edge connectivity cλg(G) of a non-complete graph G is the minimum number of edges to be deleted from the graph G such that the resulting graph has at least g components. Hypercube-like networks (HL-networks for short) are obtained by manipulating some pairs of edges in hypercubes, which contain several famous interconnection networks such as twisted cubes, Möbius cubes, crossed cubes, locally twisted cubes. In this paper, we determine the (g + 1)-component edge connectivity of the n-dimensional HL-networks for g ≤ 2[n/2], n ≥ 8. Copyright © 2021, The Authors. All rights reserved.

关键词： Graph theory

Multi-Object Localization and Irrelevant-Semantic Separation for Nuclei Segmentation in Histopathology Images

学校读者我要写书评

暂无评论

Multi-Object Localization and Irrelevant-Semantic Separation...

International Conference on Acoustics, Speech, and signal processing (ICASSP)

作者： Ya Tang Xiongjun Ye Xuanya Li Zhineng Chen Key Laboratory of Intelligent Computing and Information Processing of Ministry of Education Xiangtan University Xiangtan China National Cancer Center Cancer Hospital Chinese Academy of Medical Sciences and Peking Union Medical College Beijing China Baidu Inc. Beijing China School of Computer Science Fudan University Shanghai China

Automated segmentation of nuclei in histopathology images is critical for cancer diagnosis and prognosis. Due to the high variability of nuclei morphology, numerous nuclei overlapping, and the wide existence of nuclei clusters, this task still remains challenging. In this paper, we propose an effective nuclei segmentation method for histopathology images based on a novel neural network for multi-object localization and irrelevant-semantic separation (MI-Net), which includes a multi-object localization module (MOLM), a deep boundary awareness module (DBAM), and an irrelevant semantic separation module (ISSM). Specifically, the MOLM is used to calibrate features to capture more comprehensive information about the location of nuclei. It alleviates the problem of cell adhesion in histopathology images. The DBAM is intended to extract boundary information and solve the blurred boundary problem effectively. The ISSM is used to separate foreground features from background features and effectively address the problem of complex backgrounds in histopathology images. It also solves the low contrast problem between the target objects and the background. We evaluate MI-Net on the famous MoNuSeg dataset and compare it with ten state-of-the-art methods. MI-Net surpasses the best-performed method by margins from 1.12% to 2.29% over standard evaluation metrics, showing its effectiveness on nuclei segmentation.

关键词： Location awareness Image segmentation Histopathology Semantics Neural networks signal processing Task analysis