基于Co-Teaching的噪声标签深度学习

大规模数据在人为标记时易出现标记误差,导致数据集存在噪声标签,影响深度神经网络模型的泛化。Co-teaching等现行研究方法中的样本选择机制易使噪声样本流入被选的干净标签样本子集,在训练中难以较好地控制深度神经网络模型对被选干净样本子集的拟合。因此,文中提出一个基于Co-teaching改进的新算法。该方法通过增加两个正则化损失来分别避免模型过于信任某单一类别和陷入局部最优解中。此外,引入大学习率衰减训练方法使模型在训练初期更倾向学习干净标签样本特征以得到较好的模型参数。与Co-teaching结果相比,文中模型在20%和50%对称噪声以及45%非对称噪声环境下,在MNIST、CIFAR-10合成噪声数据集及Animal10N现实数据集上的性能均取得了提升。

Risk identification and safety assessment of human-computer interaction in integrated avionics based on STAMP

To solve the problem of risk identification and quantitative assessment for human-computer interaction(HCI)in complex avionics systems,an HCI safety analysis framework based on system-theoretical process analysis(STPA)and cognitive reliability and error analysis method(CREAM)is proposed.STPACREAM can identify unsafe control actions and find the causal path during the interaction of avionics systems and pilot with the help of formal verification tools automatically.The common performance conditions(CPC)of avionics systems in the aviation environment is established and a quantitative analysis of human failure is carried out.Taking the head-up display(HUD)system interaction process as an example,a case analysis is carried out,the layered safety control structure and formal model of the HUD interaction process are established.For the interactive behavior“Pilots approaching with HUD”,four unsafe control actions and35 causal scenarios are identified and the impact of common performance conditions at different levels on the pilot decision model are analyzed.The results show that HUD's HCI level gradually improves as the scores of CPC increase,and the quality of crew member cooperation and time sufficiency of the task is the key to its HCI.Through case analysis,it is shown that STPACREAM can quantitatively assess the hazards in HCI and identify the key factors that impact safety.

Identification and distribution of 13003 landslides in the northwest margin of Qinghai-Tibet Plateau based on human-computer interaction remote sensing interpretation

The periphery of the Qinghai-Tibet Plateau is renowned for its susceptibility to landslides.However,the northwestern margin of this region,characterised by limited human activities and challenging transportation,remains insufficiently explored concerning landslide occurrence and dispersion.With the planning and construction of the Xinjiang-Tibet Railway,a comprehensive investigation into disastrous landslides in this area is essential for effective disaster preparedness and mitigation strategies.By using the human-computer interaction interpretation approach,the authors established a landslide database encompassing 13003 landslides,collectively spanning an area of 3351.24 km^(2)(36°N-40°N,73°E-78°E).The database incorporates diverse topographical and environmental parameters,including regional elevation,slope angle,slope aspect,distance to faults,distance to roads,distance to rivers,annual precipitation,and stratum.The statistical characteristics of number and area of landslides,landslide number density(LND),and landslide area percentage(LAP)are analyzed.The authors found that a predominant concentration of landslide origins within high slope angle regions,with the highest incidence observed in intervals characterised by average slopes of 20°to 30°,maximum slope angle above 80°,along with orientations towards the north(N),northeast(NE),and southwest(SW).Additionally,elevations above 4.5 km,distance to rivers below 1 km,rainfall between 20-30 mm and 30-40 mm emerge as particularly susceptible to landslide development.The study area’s geological composition primarily comprises Mesozoic and Upper Paleozoic outcrops.Both fault and human engineering activities have different degrees of influence on landslide development.Furthermore,the significance of the landslide database,the relationship between landslide distribution and environmental factors,and the geometric and morphological characteristics of landslides are discussed.The landslide H/L ratios in the study area are mainly concentrated between 0.4 and 0.64.It means the landslides mobility in the region is relatively low,and the authors speculate that landslides in this region more possibly triggered by earthquakes or located in meizoseismal area.

Dynamic constraint and objective generation approach for real-time train rescheduling model under human-computer interaction

Real-time train rescheduling plays a vital role in railway transportation as it is crucial for maintaining punctuality and reliability in rail operations.In this paper,we propose a rescheduling model that incorporates constraints and objectives generated through human-computer interaction.This approach ensures that the model is aligned with practical requirements and daily operational tasks while facilitating iterative train rescheduling.The dispatcher’s empirical knowledge is integrated into the train rescheduling process using a human-computer interaction framework.We introduce six interfaces to dynamically construct constraints and objectives that capture human intentions.By summarizing rescheduling rules,we devise a rule-based conflict detection-resolution heuristic algorithm to effectively solve the formulated model.A series of numerical experiments are presented,demonstrating strong performance across the entire system.Furthermore,theflexibility of rescheduling is enhanced through secondary analysis-driven solutions derived from the outcomes of humancomputer interactions in the previous step.This proposed interaction method complements existing literature on rescheduling methods involving human-computer interactions.It serves as a tool to aid dispatchers in identifying more feasible solutions in accordance with their empirical rescheduling strategies.

牛津大学EMI Co-Teaching教学方法的研究与实践

在高等教育国际化发展如火如荼的今天,如何更好地开展英文媒介教学活动是高等教育工作者必须思考的重要问题.本文以牛津大学EMI(English Medium Instruction:英文媒介教学)中的Co-Teaching教学方法为研究对象,探讨如何把不同专业的教学内容通过多种教学方式协作起来,促进学生的交叉学科学习和全面发展,激发学生的创新思维,以达到提高教学效果和培养复合型人才的目的.

OPTIMIZATION DESIGN OF HYDRAULIC MANIFOLD BLOCKS BASED ON HUMAN-COMPUTER COOPERATIVE GENETIC ALGORITHM

Optimization design of hydraulic manifold blocks (HMB) is studied as acomplex solid spatial layout problem. Based on comprehensive research into structure features anddesign rules of HMB, an optimal mathematical model for this problem is presented. Usinghuman-computer cooperative genetic algorithm (GA) and its hybrid optitation strategies, integratedlayout and connection design schemes of HMB can be automatically optimized. An example is given totestify it.

STUDY ON HUMAN-COMPUTER SYSTEM FOR STABLE VIRTUAL DISASSEMBLY

The cooperative work between human being and computer based on virtual reality (VR) is investigated to plan the disassembly sequences more efficiently. A three-layer model of human-computer cooperative virtual disassembly is built, and the corresponding human-computer system for stable virtual disassembly is developed. In this system, an immersive and interactive virtual disassembly environment has been created to provide planners with a more visual working scene. For cooperative disassembly, an intelligent module of stability analysis of disassembly operations is embedded into the human-computer system to assist planners to implement disassembly tasks better. The supporting matrix for stability analysis of disassembly operations is defined and the method of stability analysis is detailed. Based on the approach, the stability of any disassembly operation can be analyzed to instruct the manual virtual disassembly. At last, a disassembly case in the virtual environment is given to prove the validity of above ideas.

Co-Teaching在英语翻译教学中的应用

本文在分析co-teaching教学模式的基础上,阐释了co-teaching在英语翻译课程教学中的应用,以期为翻译教学改革提供理论支持,提高翻译课程的教学质量。

Recent Advances on Human-Computer Dialogue

Human-Computer dialogue systems provide a natural language based interface between human and computers. They are widely demanded in network information services, intelligent accompanying robots, and so on. A Human-Computer dialogue system typically consists of three parts, namely Natural Language Understanding （NLU）, Dialogue Management （DM） and Natural Language Generation （NLG）. Each part has several different subtasks. Each subtask has been received lots of attentions, many improvements have been achieved on each subtask, respectively. But systems built in traditional pipeline way, where different subtasks are assembled sequently, suffered from some problems such as error accu- mulation and expanding, domain transferring. Therefore, researches on jointly modeling several subtasks in one part or cross different parts have been prompted greatly in recent years, especially the rapid developments on deep neural networks based joint models. There is even a few work aiming to integrate all subtasks of a dialogue system in a single model, namely end-to-end models. This paper introduces two basic frames of current dialogue systems and gives a brief survey on recent advances on variety subtasks at first, and then focuses on joint models for multiple subtasks of dialogues. We review several different joint models including integration of several subtasks inside NLU or NLG, jointly modeling cross NLG and DM, and jointly modeling through NLU, DM and NLG. Both advantages and problems of those joint models are discussed. We consider that the joint models, or end-to-end models, will be one important trend for developing Human-Computer dialogue systems.

A Study of Electromyogram Based on Human-Computer Interface

In this paper, a new control system based on forearm electromyogram （EMG） is proposed for computer peripheral control and artificial prosthesis control. This control system intends to realize the commands of six pre-defined hand poses： up, down, left, right, yes, and no. In order to research the possibility of using a unified amplifier for both electroencephalogram （EEG） and EMG, the surface forearm EMG data is acquired by a 4-channel EEG measurement system. The Bayesian classifier is used to classify the power spectral density （PSD） of the signal. The experiment result verifies that this control system can supply a high command recognition rate （average 48%） even the EMG data is collected with an EEG system just with single electrode measurement.

Design and Implementation of the Image Interactive System Based on Human-Computer Interaction

Based on the traditional Human-Computer Interaction method which is mainly touch input system, the way of capturing the movement of people by using cameras is proposed. This is a convenient technique which can provide users more experience. In the article, a new way of detecting moving things is given on the basis of development of the image processing technique. The system architecture decides that the communication should be used between two different applications. After considered, named pipe is selected from many ways of communication to make sure that video is keeping in step with the movement from the analysis of the people moving. According to a large amount of data and principal knowledge, thinking of the need of actual project, a detailed system design and realization is finished. The system consists of three important modules: detecting of the people's movement, information transition between applications and video showing in step with people's movement. The article introduces the idea of each module and technique.

Reform of Human-Computer Interaction Course based on Mobile Application Interaction Design

With the popularity of new intelligent mobile devices in people’s lives,the development of mobile applications has paid increasing attention to the interactive experience of users.As the content of traditional Human-Computer Interaction(HCI)course and teaching material is out of date,it cannot meet the needs of mobile application interaction design and enterprises for students.Therefore,we need a new generation HCI course based on intelligent mobile devices to study the relationship between users and systems.The HCI course not only teaches students HCI theory and model,but also needs to cultivate students’interaction-oriented design practical ability.This paper proposes a set of HCI teaching material design and teaching methods for improving HCI class quality on mobile application interaction design,so as to make students more suitable for the employment requirements of enterprises.

Co-teaching:The Challenges and Practice Strategies for Co-teachers

Co-teaching, an approach, provides professional instructions for students with disabilities served by a cooperation by a general education teacher and a special education teacher. This paper offers a brief overview of practical techniques and strategies drawn from different kinds of articles that are designed to support co-teachers as they are working with children and partners.

On Co-teaching and Language Immersion

The purpose of this paper is to reflect upon a pedagogic experience that was conducted by the two authors.A typical image of teaching usually points to a teacher standing out alone before a class of 20 or 30 students.In this case,we had two teachers working together in an English class with a group of 30 Chinese young boys and girls in a higher education institution in Hainan,China.This experience meant assuming a non-traditional pedagogy,which was a challenge for us considering that we had never worked together before and that the group of students was unknown to us.We tried to encourage the students to act within the concept of active learning as opposed to passive learning.And we tried to do this in an environment that could foster language immersion as much as possible.The task proved to be a challenging one.

A NEW APPROACH TO THE DESIGN OF A SYSTEM FOR FEATURE EXTRACTION BASED ON HUMAN-COMPUTER COLLABORATIVE TACTIC

We are involved in an embarrassing situation that the limited capability of automated feature extraction in digital photogrammetric systems cannot satisfy the increasing needs for rapid acquisition of semantic information for applications. Facing this challenge, a new tactic, Human-Computer Collaborative (HCC) tactic, and a corresponding new method, Operator-Object Directed (OOD) method, are proposed for the design of a system for feature extraction from large scale aerial images. We hold that in almost all technical complex systems, full automation will be neither technically feasible nor socially acceptable. The system should be designed to optimize through the cooperative operation with two agents in the system: the hurtan and the computer.

Woven fabric triboelectric nanogenerators for human-computer interaction and physical health monitoring

Triboelectric nanogenerator(TENG)converts mechanical energy into valuable electrical energy,offering a solution for future energy needs.As an indispensable part of TENG,textile TENG(T-TENG)has incredible advantages in harvesting biomechanical energy and physiological signal monitoring.However,the application of T-TENG is restricted,partly because the fabric structure parameter and structure on T-TENG performance have not been fully exploited.This study comprehensively investigates the effect of weaving structure on fabric TENGs(F-TENGs)for direct-weaving yarn TENGs and post-coating fabric TENGs.For direct-weaving F-TENGs,a single-yarn TENG(Y-TENG)with a core-sheath structure is fabricated using conductive yarn as the core layer yarn and polytetrafluoroethylene(PTFE)filaments as the sheath yarn.Twelve fabrics with five different sets of parameters were designed and investigated.For post-coating F-TENGs,fabrics with weaving structures of plain,twill,satin,and reinforced twill were fabricated and coated with conductive silver paint.Overall,the twill F-TENGs have the best electrical outputs,followed by the satin F-TENGs and plain weave F-TENGs.Besides,the increase of the Y-TENG gap spacing was demonstrated to improve the electrical output performance.Moreover,T-TENGs are demonstrated for human-computer interaction and self-powered real-time monitoring.This systematic work provides guidance for the future T-TENG’s design.

Optimization and Performance Enhancement of Gesture Recognition Algorithm Based on FMCW Millimeter-Wave Radar

Gesture recognition plays an increasingly important role as the requirements of intelligent systems for human-computer interaction methods increase.To improve the accuracy of the millimeter-wave radar gesture detection algorithm with limited computational resources,this study improves the detection performance in terms of optimized features and interference filtering.The accuracy of the algorithm is improved by refining the combination of gesture features using a self-constructed dataset,and biometric filtering is introduced to reduce the interference of inanimate object motion.Finally,experiments demonstrate the effectiveness of the proposed algorithm in both mitigating interference from inanimate objects and accurately recognizing gestures.Results show a notable 93.29%average reduction in false detections achieved through the integration of biometric filtering into the algorithm’s interpretation of target movements.Additionally,the algorithm adeptly identifies the six gestures with an average accuracy of 96.84%on embedded systems.

Virtual Keyboard:A Real-Time Hand Gesture Recognition-Based Character Input System Using LSTM and Mediapipe Holistic

In the digital age,non-touch communication technologies are reshaping human-device interactions and raising security concerns.A major challenge in current technology is the misinterpretation of gestures by sensors and cameras,often caused by environmental factors.This issue has spurred the need for advanced data processing methods to achieve more accurate gesture recognition and predictions.Our study presents a novel virtual keyboard allowing character input via distinct hand gestures,focusing on two key aspects:hand gesture recognition and character input mechanisms.We developed a novel model with LSTM and fully connected layers for enhanced sequential data processing and hand gesture recognition.We also integrated CNN,max-pooling,and dropout layers for improved spatial feature extraction.This model architecture processes both temporal and spatial aspects of hand gestures,using LSTM to extract complex patterns from frame sequences for a comprehensive understanding of input data.Our unique dataset,essential for training the model,includes 1,662 landmarks from dynamic hand gestures,33 postures,and 468 face landmarks,all captured in real-time using advanced pose estimation.The model demonstrated high accuracy,achieving 98.52%in hand gesture recognition and over 97%in character input across different scenarios.Its excellent performance in real-time testing underlines its practicality and effectiveness,marking a significant advancement in enhancing human-device interactions in the digital age.

Multiscale Feature Fusion for Gesture Recognition Using Commodity Millimeter-Wave Radar

Gestures are one of the most natural and intuitive approach for human-computer interaction.Compared with traditional camera-based or wearable sensors-based solutions,gesture recognition using the millimeter wave radar has attracted growing attention for its characteristics of contact-free,privacy-preserving and less environmentdependence.Although there have been many recent studies on hand gesture recognition,the existing hand gesture recognition methods still have recognition accuracy and generalization ability shortcomings in shortrange applications.In this paper,we present a hand gesture recognition method named multiscale feature fusion(MSFF)to accurately identify micro hand gestures.In MSFF,not only the overall action recognition of the palm but also the subtle movements of the fingers are taken into account.Specifically,we adopt hand gesture multiangle Doppler-time and gesture trajectory range-angle map multi-feature fusion to comprehensively extract hand gesture features and fuse high-level deep neural networks to make it pay more attention to subtle finger movements.We evaluate the proposed method using data collected from 10 users and our proposed solution achieves an average recognition accuracy of 99.7%.Extensive experiments on a public mmWave gesture dataset demonstrate the superior effectiveness of the proposed system.

A Survey of Gesture Recognition Using Frequency Modulated Continuous Wave Radar

With technology advances and human requirements increasing, human-computer interaction plays an important role in our daily lives. Among these interactions, gesture-based recognition offers a natural and intuitive user experience that does not require physical contact and is becoming increasingly prevalent across various fields. Gesture recognition systems based on Frequency Modulated Continuous Wave (FMCW) millimeter-wave radar are receiving widespread attention due to their ability to operate without wearable sensors, their robustness to environmental factors, and the excellent penetrative ability of radar signals. This paper first reviews the current main gesture recognition applications. Subsequently, we introduce the system of gesture recognition based on FMCW radar and provide a general framework for gesture recognition, including gesture data acquisition, data preprocessing, and classification methods. We then discuss typical applications of gesture recognition systems and summarize the performance of these systems in terms of experimental environment, signal acquisition, signal processing, and classification methods. Specifically, we focus our study on four typical gesture recognition systems, including air-writing recognition, gesture command recognition, sign language recognition, and text input recognition. Finally, this paper addresses the challenges and unresolved problems in FMCW radar-based gesture recognition and provides insights into potential future research directions.