The effect of task order on the maintainability of object-oriented software

This paper presents results from a quasi-experiment that investigates how the sequence in which maintenance tasks are performed affects the time required to perform them and the functional correctness of the changes made. Specifically, the study compares how time required and correctness are affected by (1) starting with the easiest change task and progressively performing the more difficult tasks (Easy-First), versus (2) starting with the most difficult change task and progressively performing the easier tasks (Hard-First). In both cases, the experimental tasks were performed on two alternative types of design of a Java system to assess whether the choice of the design strategy moderates the effects of task order on effort and correctness.The results show that the time spent on making the changes is not affected significantly by the task order of the maintenance tasks, regardless of the type of design. However, the correctness of the maintainability tasks is significantly higher when the task order of the change tasks is Easy-First compared to Hard-First, again regardless of design. A possible explanation for the results is that a steeper learning curve (Hard-First) causes the programmer to create software that is less maintainable overall.     

Radial Basis Function Network for Multitask Pattern Recognition

In this paper, a novel type of radial basis function network is proposed for multitask pattern recognition. We assume that recognition tasks are switched sequentially without notice to a learner and they have relatedness to some extent. We further assume that training data are given to learn one by one and they are discarded after learning. To learn a recognition system incrementally in such a multitask environment, we propose Resource Allocating Network for Multi-Task Pattern Recognition (RAN-MTPR). There are five distinguished functions in RAN-MTPR: one-pass incremental learning, task change detection, task categorization, knowledge consolidation, and knowledge transfer. The first three functions enable RAN-MTPR not only to acquire and accumulate knowledge of tasks stably but also to allocate classes to appropriate tasks unless task labels are not explicitly given. The fourth function enables RAN-MTPR to recover the failure in task categorization by minimizing the conflict in class allocation to tasks. The fifth function, knowledge transfer from one task to another, is realized by sharing the internal representation of a hidden layer with different tasks and by transferring class information of the most related task to a new task. The experimental results show that the recognition performance of RAN-MTPR is enhanced by introducing the two types of knowledge transfer and the consolidation works well to reduce the failure in task change detection and task categorization if the RBF width is properly set.     

Learner-controlled selection of tasks with different surface and structural features: Effects on transfer and efficiency

Surface task features are more salient than structural task features and thus easier to recognize for novices. It is predicted that the more salient the task features the better learners can choose personally relevant and varied tasks, which enhances learning transfer. To investigate this prediction, a 2 × 2 factorial experiment with 72 participants studied the effects of control over tasks that differ in their surface features (learner, program) and in their structural features (learner, program). Learner control over the selection of tasks with salient surface features enables learners to select personally relevant and varied tasks. This is believed to yield higher effectiveness (i.e., higher near and far transfer test performance) as well as higher efficiency (i.e., higher transfer test performance combined with lower associated mental effort). Learner control over the selection of tasks with non-salient structural features does not enable learners to select personally relevant and varied tasks and is therefore not expected to yield beneficial effects on learning. The results show positive effects of learner control over the selection of tasks with salient surface features for efficiency on the far transfer test but not for effectiveness. Theoretical and practical implications are discussed.     

基于迁移学习的讽刺检测

准确的讽刺检测对于情感分析等任务至关重要.传统的方法严重依赖于离散的人工制定的特征.现有的研究大多将讽刺检测作为一种标准的监督学习文本分类任务,但是监督学习需要有大量数据,而这些数据的收集和标注都存在困难.由于目标任务有限的数据集可能导致讽刺检测的低性能,为此将讽刺检测作为一种迁移学习任务,将讽刺标记文本的监督学习与外部分析资源的知识转移相结合.通过转移的资源知识来改进神经网络模型,以此提升对目标任务的检测性能.在公开可用的数据集上的实验结果表明,提出的基于迁移学习的讽刺检测模型优于现有较先进的讽刺检测模型.     

Haptic User Interfaces for the Visually Impaired: Implications for Haptically Enhanced Science Learning Systems

The overall quality of haptic user interfaces designed to support visually impaired students' science learning through sensorial feedback was systematically studied to investigate task performance and user behavior. Fourteen 6th- to 11th-grade students with visual impairments recruited from a state-funded blind school were asked to perform three main tasks (i.e., menu selection, structure exploration, and force recognition) using haptic user interfaces and a haptic device. This study used several dependent measures that are categorized into three types of variables: (a) task performance including success rate, workload, and task completion time; (b) user behavior defined as cursor movements proportionately represented from the user's cursor positional data; and (c) user preference. Results showed that interface type has significant effects on task performance, user behavior, and user preference, with varying degree of impact to participants with severe visual impairments performing the tasks. The results of this study as well as a set of refined design guidelines and principles should provide insights to the future research of haptic user interfaces that can be used when developing haptically enhanced science learning systems for the visually impaired.     

Utopia: A load sharing facility for large,heterogeneous distributed computer systems

Load sharing in large, heterogeneous distributed systems allows users to access vast amounts of computing resources scattered around the system and may provide substantial performance improvements to applications. We discuss the design and implementation issues in Utopia, a load sharing facility specifically built for large and heterogeneous systems. The system has no restriction on the types of tasks that can be remotely executed, involves few application changes and no operating system change, supports a high degree of transparency for remote task execution, and incurs low overhead. The algorithms for managing resource load information and task placement take advantage of the clustering nature of large-scale distributed systems; centralized algorithms are used within host clusters, and directed graph algorithms are used among the clusters to make Utopia scalable to thousands of hosts. Task placements in Utopia exploit the heterogeneous hosts and consider varying resource demands of the tasks. A range of mechanisms for remote execution is available in Utopia that provides varying degrees of transparency and efficiency. A number of applications have been developed for Utopia, ranging from a load sharing command interpreter, to parallel and distributed applications, to a distributed batch facility. For example, an enhanced Unix command interpreter allows arbitrary commands and user jobs to be executed remotely, and a parallel make facility achieves speed-ups of 15 or more by processing a collection of tasks in parallel on a number of hosts.     

Procedures for Obtaining and Testing User-Selected Terminologies

Four experiments were conducted to assess procedures for obtaining and testing user-selected terms for task-specific concepts in complex, unfamiliar word-processing instructions. Experiment 1 tested user-selected terms against both user-nominated and the original technical terms. The final three experiments employed only the user-selected and technical terms. The effect of terminology on subjects' abilities to follow the instructions was evaluated by measuring errors and task completion times during the practice period. Comprehension of the instructions was assessed by performance on a transfer task. Extensive practice produced acceptable and comparable performance for all term types. However, instruction comprehension, as measured by the transfer task, was clearly influenced by terminology. User-selected word-processing terms were more understandable than both user-nominated and the original technical terms. In addition, the present study demonstrated that transfer tasks can be more sensitive (and often more appropriate) evaluations of the goodness of a term than are learning measures.     

Design of multi-product manufacturing lines with the consideration of product change dependent inter-task times,reduced changeover and machine flexibility

Traditional research work in manufacturing line design assumes that the times between manufacturing tasks performed on a workstation are independent of the task sequence on that station. Furthermore, such studies paid little attention to the sequence dependent inter-task times in multi-model production. Since the inter-task times related to product model changeover have significant impact on manufacturing line performance, it is necessary to take into account the inter-task times explicitly in multi-model manufacturing line design. The need for this consideration is growing as product variety increases. This paper presents mathematical models of manufacturing line design with the consideration of product change related inter-task times in evaluating station times for multi-model production. An optimization model is developed using mixed integer programming to minimize manufacturing line cost. The model takes into account the recurrence of manufacturing tasks in a station to determine the machine type in a station. This paper also presents a heuristic solution procedure developed for efficient calculation. This paper also investigates how product model build sequences and inter-task times affect manufacturing line performance. The developed models will help enhance task-station assignment in multi- and mixed-model production by increasing line cost effectiveness and reducing line changeover impact as well as shortening long re-balancing processes.     

用户QoS感知的GPU集群深度学习任务动态调度

提出一种GPU集群下用户服务质量QoS感知的深度学习研发平台上的动态任务调度方法.采用离线评估模块对深度学习任务进行离线评测并构建计算性能预测模型.在线调度模块基于性能预测模型,结合任务的预期QoS,共同开展任务放置和任务执行顺序的调度.在一个分布式GPU集群实例上的实验表明,该方法相比其他基准策略能够实现更高的QoS保证率和集群资源利用率.     

Designing on‐demand education for simultaneous development of domain‐specific and self‐directed learning skills

On‐demand education enables individual learners to choose their learning pathways according to their own learning needs. They must use self‐directed learning (SDL) skills involving self‐assessment and task selection to determine appropriate pathways for learning. Learners who lack these skills must develop them because SDL skills are prerequisite to developing domain‐specific skills. This article describes the design of an on‐demand learning environment developed to enable novices to simultaneously develop their SDL and domain‐specific skills. Learners received advice on their self‐assessments and their selections of subsequent learning tasks. In the domain of system dynamics – a way to model a dynamic system and draw graphs depicting the system's behaviour over time – advice on self‐assessment is provided in a scoring rubric containing relevant performance standards. Advice on task selection indicates all relevant task aspects to be taken into account, including recommendations for suitable learning tasks which meet the individual learner's needs. This article discusses the design of the environment and the learners' perceptions of its usefulness. Most of the times, the learners found the advice appropriate and they followed it in 78% of their task selections.     

Effects of 2D wedge design as a wayfinding facilitator in a 3D virtual environment

The purpose of this study is to investigate users' wayfinding task performance and subjective preference in a three‐dimensional (3D) virtual environment. The research variables are interaction mode (touch sensitive vs body movement modes), viewing perspective [first‐person perspective (1PP) versus third‐person perspective (3PP)], and gender (man versus woman). Three difficult levels of tasks were conducted, i.e., A, A‐P, and A‐A types. The experiment is 2 × 2 × 2 between‐subject design. Participants needed to complete three wayfinding tasks and fill out subjective preference questionnaires regarding task fluency, satisfaction, system usability scale, and NASA Task Load Index (NASA‐TLX). Fifty‐six participants (28 men and 28 women) were invited using convenient sampling method. The results are as follows: (1) Participants adopting touch‐sensitive mode performed better than body movement mode. (2) Participants using touch‐sensitive mode could have better task fluency and higher SUS score, and with less effort and frustration. (3) Participants adopting 3PP had better wayfinding task performance than first‐person perspective. They also had higher system usability scale score and spent less effort for having wider viewing perspective. (4) The gender variable only showed significant effect in A‐P type of wayfinding task. Men performed better than women. It is concluded that adopting touch‐sensitive interaction and 3PP designs can better facilitate users' wayfinding tasks.     

A process simulation based method for scheduling product design change propagation

It is expected by enterprises that design changes that occur in a complex product development process can be resolved in a cost-effective way so that products modified for new customer requirements or new technology can be launched into the market without increasing much cost and time. Usually there are different solutions or methods for design changes, which may have different impacts on the whole process. Since a complex product development process consists of many design tasks and different relationships among these tasks, it is not easy to analytically compute the completion time when change propagation finishes. In this paper a process simulation based method is proposed to select the most economic propagation path for each design change, which can reduce the total process time for changes occurring in the complex product development process. “And/Or” graph nodes are used to represent the input and output logics of each task in the design process. Mathematic models are developed to calculate the completion time of the change propagation process. Monte Carlo based simulation algorithms for finding feasible output paths and calculating impacts of input changes on tasks are developed in order to simulate the change propagation process. The objective function for scheduling design change propagations is obtained by the regression analysis of simulation results. And the best change propagation paths, which usually are not the real optimum of the original design process model, are given by optimizing the regression model with a commercial optimization package. The effect of the method is demonstrated with a study of change propagations in the motor cycle engine design process.     

Effects of prior knowledge and joint attention on learning from eye movement modelling examples

Eye movement modelling examples (EMMEs) are instructional videos of a model's demonstration and explanation of a task that also show where the model is looking. EMMEs are expected to synchronize students' visual attention with the model's, leading to better learning than regular video modelling examples (MEs). However, synchronization is seldom directly tested. Moreover, recent research suggests that EMMEs might be more effective than ME for low prior knowledge learners. We therefore used a 2 × 2 between-subjects design to investigate if the effectiveness of EMMEs (EMMEs/ME) is moderated by prior knowledge (high/low, manipulated by pretraining), applying eye tracking to assess synchronization. Contrary to expectations, EMMEs did not lead to higher learning outcomes than ME, and no interaction with prior knowledge was found. Structural equation modelling shows the mechanism through which EMMEs affect learning: Seeing the model's eye movements helped learners to look faster at referenced information, which was associated with higher learning outcomes.     

Focused multi-task learning in a Gaussian process framework

Multi-task learning, learning of a set of tasks together, can improve performance in the individual learning tasks. Gaussian process models have been applied to learning a set of tasks on different data sets, by constructing joint priors for functions underlying the tasks. In these previous Gaussian process models, the setting has been symmetric in the sense that all the tasks have been assumed to be equally important, whereas in settings such as transfer learning the goal is asymmetric, to enhance performance in a target task given the other tasks. We propose a focused Gaussian process model which introduces an ??explaining away?? model for each of the additional tasks to model their non-related variation, in order to focus the transfer to the task-of-interest. This focusing helps reduce the key problem of negative transfer, which may cause performance to even decrease if the tasks are not related closely enough. In experiments, our model improves performance compared to single-task learning, symmetric multi-task learning using hierarchical Dirichlet processes, transfer learning based on predictive structure learning, and symmetric multi-task learning with Gaussian processes.     

New goals for HCI training: How to mix old and new skills in the trainee

The development of human‐computer interaction systems and the acquisition of skills associated with such systems typically occur in the context of previous experience. What is learned in one situation may facilitate or impede learning in another situation. The aim of this article is to discuss the role of experience in human‐computer interaction. The ACT? theory of skill acquisition and transfer is extended to account for the effects of old skills on the learning of new tasks. The extended model predicts a number of changes in performance that will occur when a new task involves the combination of old and new skills, including the suggestion that the learning rate of the new task will be slower than the rate at which the old skills were originally acquired. Two experiments are reported, the results of which support most of the model's predictions. The results also suggest that the minimum performance time of a task may be increased if performance of the task involves combining old and new skills. Implications of the effects of such combinations are considered with respect to the best methods of training for human‐computer interaction systems and the development of such systems.     

Task-dependent processing of tables and graphs

In two experiments participants had to detect changes in periodic sinusoidal functions, displayed in either graphic or tabular displays. Graphs had a major advantage over tables when the task required considering configurations of data. Both displays led to similar results when task performance could rely on inspecting individual data points. With graphs almost all participants reported using the optimal method for detecting changes in the function, i.e., they used the method requiring the least effort to perform the task. With tables only about half used the optimal detection method, and these participants showed transfer of learning of detection methods between tasks. Experience in using a detection method led to improved performance if the new task relied on the same method of detection. These findings demonstrate the need to consider task performance methods when determining the relative value of different displays. The set of tasks for which a display is used is likely to affect performance and needs to be analysed as a whole, since methods employed for one task can affect task performance in other tasks.     

Learning effects in the lane change task (LCT) – Realistic secondary tasks and transfer of learning

Driver distraction is a factor that is heavily involved in traffic crashes. With in-vehicle devices like navigation systems or mobile phones on the rise, the assessment of their potential to distract the driver has become a pressing issue. Several easy-to-use methods have been developed in recent years to allow for such an assessment in the early stages of product development. One of these methods is the lane change task (LCT), a simple driving simulation in which the driver has to change lanes as indicated by different signs along the road. Although the LCT is an ISO sanctioned procedure, there are still open questions. One issue are learning effects which have been found in previous studies and which have the potential to compromise the comparability of test results. In this paper, we present results on two experiments that further explored the effect of previous experience on LCT and secondary task performance. The results confirm that learning effects occur when combining the LCT with a realistic secondary task. Also, we found evidence for the transfer of learning from one secondary task to another to some degree, provided that the two tasks are sufficiently similar.     

Instructing students on effective sequences of examples and problems: Does self-regulated learning improve from knowing what works and why?

Nowadays, students often practice problem-solving skills in online learning environments with the help of examples and problems. This requires them to self-regulate their learning. It is questionable how novices self-regulate their learning from examples and problems and whether they need support. The present study investigated the open questions (1) to what extent students' (novices) task selections align with instructional design principles and (2) whether informing them about these principles would improve their task selections, learning outcomes, and motivation. Higher education students (N = 150) learned a problem-solving procedure by fixed sequences of examples and problems (FS-condition), or by self-regulated learning (SRL). The SRL participants selected tasks from a database, varying in format, complexity, and cover story, either with (ISRL-condition) or without (SRL-condition) watching a video detailing the instructional design principles. Students' task-selection patterns in both SRL conditions largely corresponded to the principles, although tasks were built up in complexity more often in the ISRL-condition than in the SRL-condition. Moreover, there was still room for improvement in students' task selections after solving practice problems. The video instruction helped students to better apply certain principles, but did not enhance learning and motivation. Finally, there were no test performance or motivational differences among conditions. Although these findings might suggest it is relatively ‘safe’ to allow students to independently start learning new problems-solving tasks using examples and problems, caution is warranted: It is unclear whether these findings generalize to other student populations, as the students participating in this study have had some experience with similar tasks or learning with examples. Moreover, as there was still room for improvement in students' task selections, follow-up research should investigate how we can further improve self-regulated learning from examples and practice problems.     

User Behavior Model Based on Affordances and Emotions: A New Approach for an Optimal Use Method in Product–User Interactions

This study proposes a new approach to developing a user behavior model to explain how a user finds the optimal use. This is achieved by considering user concerns, task significances, affordances, and emotional responses as the interaction components and by exploring behavior sequences for a goal in using a product the first time. The tasks in the same group at each level in the user concern structure are therefore in a competing relationship in going up to a higher task. The task tree with the significances and the affordance probabilities can be analyzed. The order of a user’s exploring behavior sequences can be determined through comparisons of the expected significances, which can be obtained by the modified subjective expected utility theory. A user’s emotional responses for the tasks that a behavior sequence is composed of can be calculated by the modified decision affect theory. Here, the emotional response refers to a user’s internal reactions for the degree to which a product’s affordance features can meet his or her mental model in use. The average emotional response for a behavior sequence can be a user’s decisional factor for the optimal use method in using a product with a goal. Also, the design problems of a product can be checked from users’ point of view, and the emotional losses/changes by usage failures can be discussed. For an illustrative purpose, the proposed model is applied to a numerical example with some assumptions.     

Transfer active learning by querying committee简

In real applications of inductive learning for classifi cation, labeled instances are often defi cient, and labeling them by an oracle is often expensive and time-consuming. Active learning on a single task aims to select only informative unlabeled instances for querying to improve the classifi cation accuracy while decreasing the querying cost. However, an inevitable problem in active learning is that the informative measures for selecting queries are commonly based on the initial hypotheses sampled from only a few labeled instances. In such a circumstance, the initial hypotheses are not reliable and may deviate from the true distribution underlying the target task. Consequently, the informative measures will possibly select irrelevant instances. A promising way to compensate this problem is to borrow useful knowledge from other sources with abundant labeled information, which is called transfer learning. However, a signifi cant challenge in transfer learning is how to measure the similarity between the source and the target tasks. One needs to be aware of different distributions or label assignments from unrelated source tasks;otherwise, they will lead to degenerated performance while transferring. Also, how to design an effective strategy to avoid selecting irrelevant samples to query is still an open question. To tackle these issues, we propose a hybrid algorithm for active learning with the help of transfer learning by adopting a divergence measure to alleviate the negative transfer caused by distribution differences. To avoid querying irrelevant instances, we also present an adaptive strategy which could eliminate unnecessary instances in the input space and models in the model space. Extensive experiments on both the synthetic and the real data sets show that the proposed algorithm is able to query fewer instances with a higher accuracy and that it converges faster than the state-of-the-art methods.