检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Meng, Lingchen Yang, Jianwei Tian, Rui Dai, Xiyang Wu, Zuxuan Gao, Jianfeng Jiang, Yu-Gang Shanghai Key Lab of Intell. Info. Processing School of CS Fudan University China Shanghai Collaborative Innovation Center of Intelligent Visual Computing China Microsoft Corporation United States

Most large multimodal models (LMMs) are implemented by feeding visual tokens as a sequence into the first layer of a large language model (LLM). The resulting architecture is simple but significantly increases computation and memory costs, as it has to handle a large number of additional tokens in its input layer. This paper presents a new architecture DeepStack for LMMs. Considering N layers in the language and vision transformer of LMMs, we stack the visual tokens into N groups and feed each group to its aligned transformer layer from bottom to top, as illustrated in Fig. 1. Surprisingly, this simple method greatly enhances the power of LMMs to model interactions among visual tokens across layers but with minimal additional cost. We apply DeepStack to both language and vision transformer in LMMs, and validate the effectiveness of DeepStack LMMs with extensive empirical results. Using the same context length, our DeepStack 7B and 13B parameters surpass their counterparts by 2.7 and 2.9 on average across 9 benchmarks, respectively. Using only one-fifth of the context length, DeepStack rivals closely to the counterparts Copyright © 2024, The Authors. All rights reserved.

关键词： Visual languages

来源：评论

学校读者我要写书评

暂无评论

NetPrompt: Neural Network Prompting Enhances Event Extraction in Large Language Models

引用

IEEE Transactions on Big Data 2025年

作者： Mu, Lin Cheng, Yide Shen, Jun Zhang, Yiwen Zhong, Hong School of Computer Science and Technology Anhui University Anhui Hefei230601 China The Key Laboratory of Intelligent Computing and Signal Processing of Ministry of Education the Anhui Engineering Laboratory of IoT Security Technologies the School of Computer Science and Technology the Institute of Physical Science and Information Technology Anhui University Hefei230039 China

Event Extraction involves extracting event-related information such as event types and event arguments from context, which has long been tackled through well-designed neural networks or fine-tuned pre-trained language models. These approaches require substantial annotated data for tuning parameters and are resource-intensive. Recently, Prompting strategies with frozen parameters, such as Chain-of-Thought and Self-Consistency, have delivered success in NLP using LLMs by generating intermediate thought steps. However, they suffer from the challenge of error propagation and lack of interaction between different thoughts. In this paper, we propose Neural Network-based Prompting (NetPrompt), a novel network-structured prompting strategy for event extraction. The core idea behind NetPrompt is to imitate the excellent information integration capabilities of neural network structures. Specifically, we first decompose the event extraction problem into diverse intermediate subtasks, and each subtask is represented as a node in different layers of the network, the output of the nodes in the preceding layer is fed into the subsequent layer. Secondly, we propose pruning strategies to adapt the reasoning overhead to different problems. Finally, we have conducted extensive experiments on two widely used event extraction benchmarks to evaluate NetPrompt. The results demonstrated that NetPrompt significantly improved the event extraction performance compared to previous methods. ©2015 IEEE.

关键词： Neural network models

来源：评论

学校读者我要写书评

暂无评论

A dual-band antenna designed for 5G communication based on a self-decoupled antenna

A dual-band antenna designed for 5G communication based on a...

引用

International Applied Computational Electromagnetics Society Symposium (ACES)

作者： Rongzhen Fang Xianliang Wu Zhipeng Chen Mengying Dong Yunfeng Hu Yuanyuan Li Key Laboratory of Intelligent Computing and Signal Processing Anhui University Hefei China Department of Education of Anhui Province Anhui University Hefei China

In this paper, a dual-band antenna for 5G communication based on an antenna design with self-decoupling properties is *** antenna is composed of a self-decoupled antenna unit vertically placed on an 30×80 mm 2 ground plane and printed on 0.8 mm thick FR-4(εr = 4.4, tanδ = 0.02) substrate,and then added a pair of L-shaped coupling feeder structures on this *** shows that the antenna has two operating frequency bands of 3.3-4.2GHz and 4.8-5GHz, and has good transmission and isolation in the above two operating frequency ***,it also has the advantages of small size, self-decoupling,high isolation,simple structure and easy *** antenna can be used as 5G mobile phone communication antenna unit.

关键词： Couplings 5G mobile communication Transmitting antennas Dual band Mobile antennas Antenna feeds Computational electromagnetics

来源：评论

学校读者我要写书评

暂无评论

Electromagnetic scattering of two-dimensional rough cylinder based on statistical integral equation 2

Electromagnetic scattering of two-dimensional rough cylinder...

引用

2020 2nd International Conference on Artificial Intelligence Technologies and Application, ICAITA 2020

作者： Zhu, Mengnan Wang, Anqi Huang, Zhixiang Key Laboratory of Intelligent Computing and Signal Processing Ministry of Education Anhui University Hefei China

The study of electromagnetic scattering from Gaussian rough surface is of great significance in radar reconnaissance, target tracking and ocean remote sensing. The moment method (MOM) is a commonly used method with high accuracy. However, in the past research, Monte Carlo method is used to simulate the Gaussian rough surface. In this method, a large number of random simulation operations are carried out, and then a statistical convergence result is obtained by counting the results of all operations. In this paper, the stochastic integral equation method SIEM(Stochastic Integral Equation Method) is proposed to calculate the electromagnetic scattering of Gaussian rough surface. SIEM avoids the difficulty of high difficulty precise modeling and repeated calculation many times, and only single modeling and calculation are carried out for the target problem. In this paper, the electromagnetic scattering of two-dimensional cylinder with Gaussian rough surface is solved, and the effectiveness of SIEM is proved by comparing with mom. © 2020 Published under licence by IOP Publishing Ltd.

关键词： Method of moments

来源：评论

学校读者我要写书评

暂无评论

Expanding self-orthogonal codes over a ring Z4 to self-dual codes and unimodular lattices

arXiv

引用

arXiv 2024年

作者： Shi, Minjia Tao, Sihui Hong, Jihoon Kim, Jon-Lark The Key Laboratory of Intelligent Computing Signal Processing Ministry of Education School of Mathematical Sciences Anhui University Anhui Hefei230601 China State Key Laboratory of Integrated Service Networks Xidian University Xi’an710071 China The Department of Mathematics Sogang University Seoul Korea Republic of

Self-dual codes have been studied actively because they are connected with mathematical structures including block designs and lattices and have practical applications in quantum error-correcting codes and secret sharing schemes. Nevertheless, there has been less attention to construct self-dual codes from self-orthogonal codes with smaller dimensions. Hence, the main purpose of this paper is to propose a way to expand any self-orthogonal code over a ring Z4 to many self-dual codes over Z4. We show that all self-dual codes over Z4 of lengths 4 to 8 can be constructed this way. Furthermore, we have found five new self-dual codes over Z4 of lengths 27, 28, 29, 33, and 34 with the highest Euclidean weight 12. Moreover, using Construction A applied to our new Euclidean-optimal self-dual codes over Z4, we have constructed a new odd extremal unimodular lattice in dimension 34 whose kissing number was not previously known. © 2024, CC0.

关键词： Crystal lattices

来源：评论

学校读者我要写书评

暂无评论

A RIS-Assisted Multiuser MIMO Communication Method Based on Deep Reinforcement Learning

A RIS-Assisted Multiuser MIMO Communication Method Based on ...

引用

IEEE International Conference on Communication and Information Systems (ICCIS)

作者： Changlin Tian Hui Zhi Kui Tie Key Laboratory of Intelligent Computing and Signal Processing of Ministry of Education Anhui University School of Electronic and Information Engineering Anhui University Hefei China Ceyear Technologies (AnHui) CO. LTD Bengbu China

In this paper, a RIS-assisted multiuser MIMO communication method based on deep reinforcement learning (RMMC-DRL) is proposed for multiuser scenarios. Our objective is to find the optimal transmit beamforming matrix of BS and optimal phase shift matrix of reflective intelligent surface (RIS) to maximize the sum rate of multiuser, this problem is reduced into a constrained optimization problem. It is a non-convex optimization problem, so we solve it through deep reinforcement learning (DRL) and then use the results for communication. In the DRL, a deep deterministic policy gradient (DDPG) framework that can handle continuous states and actions is designed, reward is set as optimization goal, and the transmit beamforming matrix and the phase shift matrix of RIS are obtained through the interaction with environment. Unlike the alternating optimization (AO) method, which solve the transmit beamforming matrix and the RIS phase shift matrix alternatively, the RMMC-DRL can obtain both transmit beamforming matrix and RIS phase shift matrix simultaneously as the output of DRL. Simulation results show that RMMC-DRL can learn and improve its behavior by interacting with the environment. Compared with AO method, RMMC-DRL can obtain higher sum rate and lower computational complexity.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A 0.33–2.61-GHz Rectifier with Expanded Dynamic Input Power Range Using Microstrip Impedance Compression Circuit

TechRxiv

引用

TechRxiv 2023年

作者： Li, Yingsong Liu, Xin Zhang, Dawei Wu, Xiaohu Liu, Xiaoguang Huang, Zhixiang The College of Information and Communication Engineering Harbin Engineering University Harbin150001 China The Key Laboratory of Intelligent Computing and Signal Processing Ministry of Education Anhui University Anhui Hefei China Shenzhen518055 China

In this letter, an ultra-broadband rectifier with expanded dynamic input power range (IPR) for both wireless power transfer (WPT) and radio frequency energy-harvesting (RFEH) is proposed and analyzed. Expanded dynamic IPR and broadband impedance matching are realized by utilizing a microstrip transmission line impedance compression circuit (MTLICC) with the topology of the paralleled voltage doubler. A proof-of-concept prototype shows a dynamic IPR of 8.5–23.5 dBm, a frequency range of 0.33–2.61 GHz (fractional bandwidth of 155%), PCE is greater than 50%, and a peak PEC of 79.7% (achieved at an input-power level of 19 dBm). The measurement results are in good agreement with the simulation. Covering most 3G/4G/5G/ISM frequency bands, this design is promising for RF power transmission and energy harvesting in the age of the Internet of Things systems. © 2023, CC BY.

关键词： Timing circuits

来源：评论

学校读者我要写书评

暂无评论

A Unified Affine-Projection-Like Adaptive Algorithm for System Identification

TechRxiv

引用

TechRxiv 2023年

作者： Li, Yingsong Fu, Yonglin Miao, Yongchun Huang, Zhixiang Diniz, Paulo S.R. The Key Laboratory of Intelligent Computing and Signal Processing Ministry of Education Anhui University Anhui Hefei230601 China The Universidade Federal do Rio de Janeiro Rio de Janeiro21941-972 Brazil

A unified affine-projection-like adaptive (UAPLA) algorithm is deivised and verified for system identification. The UAPLA algorithm uses a generalized cost function encompassing some data-reusing methods to cope with colored input signals. Furthermore, the UAPLA algorithm is derived based on the new cost function to generalize multiple affine-projection algorithms. As a result, the proposed UAPLA algorithm includes some classical adaptive filters as exceptional cases while allowing flexibility to achieve low estimation errors under impulsive noises. The obtained results conducted by simulations help to corroborate the superiority of the proposed UAPLA algorithm over other popular AP algorithms. © 2023, CC BY.

关键词： Adaptive filters

来源：评论

学校读者我要写书评

暂无评论

Top-ambiguity samples matter: understanding why deep ensemble works in selective classification 23

Top-ambiguity samples matter: understanding why deep ensembl...

引用

Proceedings of the 37th International Conference on Neural Information processing Systems

作者： Qiang Ding Yixuan Cao Ping Luo Key Lab of Intelligent Information Processing of Chinese Academy of Sciences (CAS) Institute of Computing Technology CAS Beijing China and University of Chinese Academy of Sciences Beijing China Key Lab of Intelligent Information Processing of Chinese Academy of Sciences (CAS) Institute of Computing Technology CAS Beijing China and University of Chinese Academy of Sciences Beijing China and Peng Cheng Laboratory Shenzhen China

Selective classification allows a machine learning model to reject some hard inputs and thus improve the reliability of its predictions. In this area, the ensemble method is powerful in practice, but there has been no solid analysis on why the ensemble method works. Inspired by an interesting empirical result that the improvement of the ensemble largely comes from top-ambiguity samples where its member models diverge, we prove that, based on some assumptions, the ensemble has a lower selective risk than the member model for any coverage within a range. The proof is nontrivial since the selective risk is a non-convex function of the model prediction. The assumptions and the theoretical results are supported by systematic experiments on both computer vision and natural language processing tasks.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Pix2Cap-COCO: Advancing Visual Comprehension via Pixel-Level Captioning

arXiv

引用

arXiv 2025年

作者： You, Zuyao Wang, Junke Kong, Lingyu He, Bo Wu, Zuxuan Shanghai Key Lab of Intell. Info. Processing School of CS Fudan University China Shanghai Collaborative Innovation Center of Intelligent Visual Computing China University of Maryland College Park United States

(Image Present) We present Pix2Cap-COCO, the first panoptic pixel-level caption dataset designed to advance fine-grained visual understanding. To achieve this, we carefully design an automated annotation pipeline that prompts GPT-4V to generate pixel-aligned, instance-specific captions for individual objects within images, enabling models to learn more granular relationships between objects and their contexts. This approach results in 167,254 detailed captions, with an average of 22.94 words per caption. Building on Pix2Cap-COCO, we introduce a novel task, panoptic segmentation-captioning, which challenges models to recognize instances in an image and provide detailed descriptions for each simultaneously. To benchmark this task, we design a robust baseline based on X-Decoder. The experimental results demonstrate that Pix2Cap-COCO is a particularly challenging dataset, as it requires models to excel in both fine-grained visual understanding and detailed language generation. Furthermore, we leverage Pix2Cap-COCO for Supervised Fine-Tuning (SFT) on large multimodal models (LMMs) to enhance their performance. For example, training with Pix2Cap-COCO significantly improves the performance of GPT4RoI, yielding gains in CIDEr (+1.4%), ROUGE (+0.4%), and SPICE (+0.5%) on Visual Genome dataset, and strengthens its region understanding ability on the ViP-Bench, with an overall improvement of +5.1%, including notable increases in recognition accuracy (+11.2%) and language generation quality (+22.2%). Code is available at https://***/geshang777/pix2cap. © 2025, CC BY-NC-SA.

关键词： Pixels

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：