检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Qi, Biqing Li, Pengfei Li, Fangyuan Gao, Junqi Zhang, Kaiyan Zhou, Bowen Department of Control Science and Engineering Harbin Institute of Technology China Department of Electronic Engineering Tsinghua University China School of Mathematics Harbin Institute of Technology China

Direct Preference Optimization (DPO) improves the alignment of large language models (LLMs) with human values by training directly on human preference datasets, eliminating the need for reward models. However, due to the presence of cross-domain human preferences, direct continual training can lead to catastrophic forgetting, limiting DPO’s performance and efficiency. Inspired by intraspecific competition driving species evolution, we propose a Online Fast-Slow chasing DPO (OFS-DPO) for preference alignment, simulating competition through fast and slow chasing among models to facilitate rapid adaptation. Specifically, we first derive the regret upper bound for online learning, validating our motivation with a min-max optimization pattern. Based on this, we introduce two identical modules using Low-rank Adaptive (LoRA) with different optimization speeds to simulate intraspecific competition, and propose a new regularization term to guide their learning. To further mitigate catastrophic forgetting in cross-domain scenarios, we extend the OFS-DPO with LoRA modules combination strategy, resulting in the Cross domain Online Fast-Slow chasing DPO (COFS-DPO). This method leverages linear combinations of fast modules parameters from different task domains, fully utilizing historical information to achive continual value alignment. Experimental results show that OFS-DPO outperforms DPO in in-domain alignment, while COFS-DPO excels in cross-domain continual learning scenarios. Copyright © 2024, The Authors. All rights reserved.

关键词： Alignment

来源：评论

学校读者我要写书评

暂无评论

Research on an Intelligent Routing Strategy for Industrial Internet of Things 2nd

Research on an Intelligent Routing Strategy for Industrial I...

引用

2nd EAI International Conference on Artificial Intelligence for Communications and Networks, AICON 2020

作者： Zhang, Xu Department of Control Science and Engineering Harbin Institute of Technology Harbin150080 China

ISBN: (纸本)9783030690656

The Industrial Internet of Things is considered to be an important cornerstone of future industrial development and has broad development prospects. It is being deployed to the society on a large scale. Sensor nodes in the Internet of Things will inevitably face many challenges. Due to the limitations of the existing sensor nodes, the traditional energy measurement methods for wireless sensor networks have been difficult to meet. Therefore, this article proposes a new routing protocol for low-power and low-power networks (RPL). The routing measurement method EEM meets the requirements of low-power lossy networks for link quality and energy consumption, and then uses the modified simulation software to test the EEM routing measurement, and performs performance verification on the packet loss rate and network load. The experimental results show that EEM retains the requirements of ETX routing metrics for link quality, realizes the awareness of node energy consumption, and optimizes network load. © 2021, ICST inst.tute for Computer sciences, Social Informatics and Telecommunications engineering.

关键词： Internet of things

来源：评论

学校读者我要写书评

暂无评论

Intelligent photothermal treatment method based on photoacoustic and photothermal probe assistance 4

Intelligent photothermal treatment method based on photoacou...

引用

4th Optics Young Scientist Summit, OYSS 2020

作者： Ma, Yiming Ma, Liyong Gao, Yuanyuan Sun, Mingjian Department of Control Science and Engineering Harbin Institute of Technology Harbin150000 China

ISBN: (数字)9781510643970

ISBN: (纸本)9781510643963

As a non-invasive, targeted and non-radioactive technology applied to tumor treatment, photothermal therapy is increasingly used in the clinical treatment of tumors due to its high cure rate and few side effects. In order to obtain a better photothermal treatment effect in the actual treatment process, the temperature distribution of the tissue to be treated must be monitored and controlled to prevent unnecessary tissue damage caused by the treatment. Therefore, the photothermal probe and precise control of the treatment temperature become the key to solving the problem. In this paper, a nanoprobe with strong photoacoustic and photothermal properties in one area of near-infrared is designed. At the same time, a photothermal treatment system is designed and combined with nanoprobe for research. This study designed human tissue simulation experiments and found that compared to the case without probe assistance, the photothermal therapy system based on photoacoustic and photothermal probe assistance can achieve temperature control with a temperature error of 1°C, and the temperature control adjustment time has been shortened by 40%, and the overtreatment injury has also been effectively suppressed. More importantly, it is possible to greatly reduce the complexity of the photothermal treatment process in practical applications without the artificial control of continuous laser power. The experimental result shows that the intelligent photothermal treatment method based on the photoacoustic and photothermal nanoprobe is a feasible candidate for tumor treatment and has a good application prospect in the field of tumor treatment. © SPIE. Downloading of the abstract is permitted for personal use only.

关键词： Nanoprobes

来源：评论

学校读者我要写书评

暂无评论

Stabilization with Prescribed inst.nt via Lyapunov Method

arXiv

引用

arXiv 2023年

作者： Kuang, Jiyuan Gao, Yabin Sun, Yizhuo Wang, Jiahui Liu, Aohua Zhao, Yue Liu, Jianxing The Department of Control Science and Engineering Harbin Institute of Technology Harbin150001 China The College of Intelligent System Science and Engineering Harbin Engineering University Harbin150001 China

This letter investigates the prescribed-inst.nt stabilization problem for high-order integrator systems. In anothor word, the settling time under the presented controller is independent of the initial conditions and equals the prescribed time inst.nt. The controller is designed with the concept of backstepping. A strict proof based on the Lyapunov method is presented to clamp the settling time to the prescribed time inst.nt from both the left and right sides. This proof serves as an example to present a general framework to verify the designed stabilization property. It should be emphasized that the prescribed-time stability (PSTS) [1] can only prescribe the upper bound of the settling time and is different from this work. The detailed argumentation will be presented after a brief review of the existing important research. © 2023, CC BY.

关键词： Lyapunov methods

来源：评论

学校读者我要写书评

暂无评论

An Improved Classification Method of Ultrasonic Thyroid Standard Planes Based on Bayesian Optimization of Multi-Feature Parameters for Portable Application

An Improved Classification Method of Ultrasonic Thyroid Stan...

引用

Chinese control Conference (CCC)

作者： Xinran Li Xin Zhang Chenyang Fan Yi Shen Department of Control Science and Engineering Harbin Institute of Technology Harbin PR China School of Mathematics Harbin Institute of Technology Harbin PR China

In recent years, computer aided technology is increasingly applied to the classification of thyroid ultrasonic planes for calculating the thyroid volume, and then the disease can be diagnosed by its volume. However, these methods requiring greater computational power are not suitable for the portable ultrasound system with limited performance. Based on Bayesian optimization of multi-feature parameters, an improved classification method of ultrasonic thyroid planes is proposed in this paper for low computational complexity with a high classification accuracy. Local Binary Patterns (LBP) feature and Gray Level Co-occurrence Matrix (GLCM) feature are selected by the analysis of image texture information, and then the combined feature is utilized to investigate the effectiveness of different classifiers. Furthermore, Bayesian optimization is employed to obtain the optimal parameters of the combined feature for improving classification results, and the accuracy can reach up to 96.81%. The results clearly illustrate that the improved method is effective in classifying the ultrasonic thyroid planes under a low computational complexity.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Coverage control for Multi-Agent Systems via Curve Arc Length

Coverage Control for Multi-Agent Systems via Curve Arc Lengt...

引用

Chinese Automation Congress (CAC)

作者： Kaiqi Yang Xiaoli Wang Hengyu Yang Wei Xie Department of Control Science and Engineering Harbin Institute of Technology Harbin China School of Information Science and Engineering Harbin Institute of Technology at Weihai Weihai China School of Aeronautics and Astronautics Tiangong University Tianjin China

ISBN: (数字)9798350368604

ISBN: (纸本)9798350368611

In complex environments, effective and comprehensive information gathering necessitates the strategic deployment of multi-agent systems. Corridor-type regions, characterized as non-convex areas, present substantial challenges in formulating coverage control strategies. To tackle this issue, this paper proposes a novel coverage control methodology for multi-agent systems based on partitioning the coverage area using curve arc lengths. Firstly, the coverage areas for each team is segmented into sequentially arranged subregions based on curve arc lengths. Through adjustments in the virtual leading agent's position, the range of curve arc lengths for each team's coverage areas is dynamically managed. Then, a distributed coverage controller is developed for agents, which integrates the movement status of virtual leading agents to guide agents towards the centroid positions of their respective coverage areas. Finally, numerical simulations affirm the efficacy of the proposed coverage control approach.

关键词： Automation control systems Numerical simulation Multi-agent systems

来源：评论

学校读者我要写书评

暂无评论

Autotuned Non-Linear Extended-State-Observer Based Fixed-Time control for Motor-Drive Servo Systems

引用

IEEE Transactions on Power Electronics 2025年

作者： Xu, Ruiqi Liu, Jianxing Lin, Xinpo Liu, Zhuang Yan, Fei Gao, Yabin Harbin Institute of Technology Department of Control Science and Engineering Harbin150001 China Harbin Institute of Technology Department of Control Science and Engineering Weihai264209 China Southwest Jiaotong University School of Information Science and Technology Chengdu611756 China

To enhance the position tracking performance of permanent magnet synchronous motors (PMSMs) under unknown disturbances, this paper proposes a fast fixed-time control strategy based on an improved nonlinear extended state observer (NESO). First, load disturbances and nonlinear friction are considered as lumped disturbance, for which a NESO is designed to estimate the lumped disturbance. The control parameters of the NESO are tuned online using a neural network optimization algorithm, eliminating the need for offline training. Then, a robust fixed-time sliding mode control method is proposed, based on an improved nonsingular fast terminal sliding mode manifold, which offers better convergence performance. The Lyapunov method is used to prove the fixed-time stability of the position tracking error system. Finally, the effectiveness of the proposed method is validated on an experimental platform with PMSMs, and it is compared with other advanced fixed-time position control methods. The comparison results confirm that the proposed method exhibits superior control performance. © 1986-2012 IEEE.

关键词： State estimation

来源：评论

学校读者我要写书评

暂无评论

High-Performing Honeycomb-Structured Triboelectric Nanogenerator Enhanced by Triple Electrodes for Utilizing Wind Power

SSRN

引用

SSRN 2023年

作者： Liu, Xu Liu, Yizhi Liu, Fante Zhang, Zidi Zhang, Zeyu Yang, Zhiqiang Department of Astronautical Science and Mechanics Harbin Institute of Technology Harbin China Department of Control Science and Engineering School of Astronautics Harbin Institute of Technology China

The utilization of wind power and other renewable resources is attracting more and more attention worldwide in response to climate change. However, the majority of energy-converting facilities are usually frail and susceptible to natural factors, which leads to inefficient and inconsistent operation. Here, originating from bionics and nanotechnology, the honeycomb-structured triboelectric nanogenerator with triple-electrode is proposed, permitting long-term efficient and stable working cycles. A simple power management strategy is implemented to enhance the output current by 42.1%. The overall structure exhibits high power density (291.5 W/m3) and motivating ability under low rotational speed (60 rpm), excellent force resistance and specific energy absorption (4.03×104 J/m3), as well as flexible integration. Further, a wireless wind speed detection system is managed to monitor the changeable wind conditions intelligently. This work illuminates a feasible and effective solution for converting wind power into electricity on a large scale, providing a feasible solution to the energy crisis. © 2023, The Authors. All rights reserved.

关键词： Electrodes

来源：评论

学校读者我要写书评

暂无评论

Hierarchical carbon nanotube-decorated polyacrylonitrile smart textiles for wearable biomonitoring

Wearable Electronics

引用

Wearable Electronics 2024年 1卷 180-188页

作者： Junlong Huang Yulin Cai Guangzhong Xie Xiangdong Xu Zhenya Geng Yadong Jiang Yuanjie Su State Key Laboratory of Electronic Thin Films and Integrated Devices School of Optoelectronic Science and Engineering University of Electronic Science and Technology of China (UESTC) Chengdu 610054 China Department of Control Science and Engineering Harbin Institute of Technology Harbin 150001 China

Respiration monitoring enables continuous assessment of physiological status and potential diseases. We reported a polyacrylonitrile/carbon nanotubes/latex composite membrane (PCM) capable of converting exhalation into a distinct current signal for respiratory rate and depth detection. The latex encapsulation design of respiratory sensor screens moist and thermal fluctuation of respiration and enables an accurate and reliable monitoring of exhaled gas flow. By modulating the compositional ratio, topological characters of interdigital electrode and aperture area, a rapid response time of 192 ms and recovery time of 104 ms, together with great fidelity and reliability were attained for discriminating breathing depth and rate. Combining with machine learning, versatile breathing patterns can be effectively discerned, achieving effective prognosis of respiratory diseases like diabetes, wheezing and obstructive sleep apnoea syndrome. This work endows a low-cost, efficient and stable route for sustainable wearable physiological assessment and disease recognition.

关键词： Respiratory sensor Hierarchical-structured PAN/CNT composite membrane Physiological monitoring Machine learning

来源：评论

学校读者我要写书评

暂无评论

Globally Composite-Learning-Based Intelligent Fast Finite-Time control for Uncertain Strict-Feedback Systems with Nonlinearly Periodic Disturbances

arXiv

引用

arXiv 2023年

作者： Wang, Xidong Li, Zhan He, Zhen The Research Institute of Intelligent Control and Systems Harbin Institute of Technology Harbin150001 China The Department of Control Science and Engineering Harbin Institute of Technology Harbin150001 China

This brief aims at the issue of globally composite-learning-based neural fast finite-time (F-FnT) tracking control for a class of uncertain systems in strict-feedback form subject to nonlinearly periodic disturbances. First, uncertain dynamics with periodic parameters are identified by incorporating Fourier series expansion (FSE) into an intelligent estimator, which leverages the feedback of newly designed prediction errors in updating weights to boost learning performance. Then, a novel switching mechanism is constructed to fulfill smooth switching from the composite FSE-based neural controller to robust control law when the inputs of the intelligent estimator transcend the valid approximation domain. By fusing the switching mechanism with an improved F-FnT backstepping algorithm, the globally F-FnT boundedness of all variables in the closed-loop system is guaranteed. Finally, a simulation study is conducted to evince the availability of the theoretical result. Copyright © 2023, The Authors. All rights reserved.

关键词： Backstepping

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：