Search

This paper describes a strategy to build community capacity to address the challenge of SDG#7 (Affordable and Clean Energy) through an international student service-learning programme. The Hong Kong Polytechnic University partnered with an NGO in Rwanda in a four-year collaborative effort to electrify a semi-rural village in Rwanda with solar power. For sustainability, we developed a three-tier capacity building model that combined efforts from university students (who provide the technical expertise), strengths and expertise of a partnering local NGO (who had previously developed a community support infrastructure in the form of strong self-help groups), and community commitment and support (who provide the manpower initially and subsequently acquired the technical capacity).

Service-learning has been proven to be a high-impact educational pedagogy in many disciplines. It addresses human, and community needs through engaging in community activities. With professional engineering expertise, integrating engineering into service-learning not only can make particularly impactful community service, especially in developing countries where engineering expertise is not always available but also provides an effective way for students to apply their theoretical knowledge to solve real-world problems. Service-learning in engineering has been documented in the last 20 years, and it was implemented as extracurricular activities followed by integrating into core curriculums. However, most of the programs are only offer to the engineering students and less consider implementing as a form of general education with applying multidisciplinary approach. This research examines how to integrate students from very different disciplines into the same project. Our case study is a credit-bearing service-learning subject offered by the Department of Computing which is open to all undergraduate students and the projects were conducted in Rwanda and Cambodia in 2015 and 2016.

This paper presents an investigation into user intention prediction in two common web-based tasks: crowdsourcing annotation and web search, based on human-mouse interaction information. User experience is gaining importance within the research area of human-centered computing, and is particularly useful for complex, multi-step tasks. To enhance user experience, the computer should be intelligent enough to be able to predict the user intention. For instance, an intelligent agent might be able to anticipate when the user is about to press a button, and helpfully enlarge or highlight it in advance. In this paper, we propose two prediction models on user intention: a classical model that considers only historical mouse activity sequence, and a multimodal model that utilizes mouse interaction signals as well as features extracted from mouse trajectory and clicking events. We evaluate our models and find that they achieve reasonable accuracy. Our preliminary results indicate that we can dynamically learn a multimodal model that can effectively predict a user’s next activity from historical activity sequence and mouse interaction signals.

The use of Information and Communication Technology (ICT) in K-12 education has been gaining attention in recent years. With using appropriate technology, not only can it allow people to have equal opportunity in education, but also improve the quality of teaching and learning. However, this is difficult in developing countries due to the lack of resources, infrastructure, support from the local community, skills, culture and policy. In 2014, according to World DataBank indicators, over half of the population in East Asia & Pacific region were not able to access the Internet. In Cambodia, this figure drops to less than one-tenth. This paper investigates complementing education in developing countries through appropriate technology and a sustainable model that maximizes short-term efforts from foreign teams. We present a case study in Cambodia whereby a zero-carbon community learning centre was designed and installed for a rural community. The first phase deployment involved a recycled twenty-foot container, furnished with solar panels, computer devices, Internet hotspot, books and toys. A sustainable model was developed through engaging local community leaders and empowering resident university students. Follow-up evaluations suggest that community ownership and buy-in are critical for success. We present design principles, implementation methods, evaluations from the experience, feedback from stakeholders, and lessons learnt.

Systems that adapt to changes in human needs automatically are useful, built upon advancements in human-computer interaction research. In this paper, we investigate the problem of how well the eye movement of a user when reading an article can predict the level of reading comprehension, which could be exploited in intelligent adaptive e-learning systems. We characterize the eye movement pattern in the form of eye gaze signal. We invite human subjects in reading articles of different difficulty levels being induced to different comprehension levels. Machine-learning techniques are applied to identify useful features to recognize when readers are experiencing difficulties in understanding their reading material. Finally, a detection model that can identify different levels of user comprehension is built. We achieve a performance improvement of over 30% above the baseline, translating over 50% reduction in detection error.

Social signal processing is becoming an important topic in affective computing. In this paper, we focus on an important social interaction in real life, namely, fighting. Fight detection will be useful in public transportation, prisons, bars, or even sport. A robust mechanism in detecting fights from a video will be extremely useful, especially in applications relevant to surveillance systems. Recent research works focus on extracting visual features from high resolution video, leading to computationally expensive systems. In this paper, we propose an approach to detect fights in a natural and robust way based on motion analysis, which is not only intuitive, but also robust. Experimental results show that we can accurately detect fight activities in different video surveillance settings.

Human-centered computing is rapidly becoming a major research direction in human-computer interaction research. Among the various research issues, we believe that affective computing, or the ability of computers to react according to what a user feels, is very important. In order to recognize the human affect (feeling), one can rely on the analysis of signal inputs captured by a multitude of means. In this paper, we propose to make use of human physiological signals as a new form of modality in determining human affects, in a non-intrusive manner. This is achieved via the physiological mouse, as a first step towards affective computing. We augment the mouse with a small optical component for capturing user photoplethysmographic (PPG) signal. With the PPG signal, we are able to compute and derive human physiological signals. We built a prototype of the physiological mouse and measured raw PPG readings. We performed experiments to study the accuracy of our approach in determining human physiological signals from the mouse PPG data. We believe that our research will provide a new dimension for multimodal affective computing research.

Most of the existing recommender systems for tourism apply knowledge-based and content-based approaches, which need sufficient historical rating information or extra knowledge and suffer from the cold start problem. In this paper, a demographic recommender system is utilized for the recommendation of attractions. This system categorizes the tourists using their demographic information and then makes recommendations based on demographic classes. Its advantage is that the history of ratings and extra knowledge are not needed, so a new tourist can obtain recommendation. Focusing on the attractions on Trip Advisor, we use different machine learning methods to produce prediction of ratings, so as to determine whether these approaches and demographic information of tourists are suitable for providing recommendations. Our preliminary results show that the methods and demographic information can be used to predict tourists' ratings on attractions. But using demographic information alone can only achieve limited accuracy. More information such as textual reviews is required to improve the accuracy of the recommendation.

Multimodality inputs are becoming increasingly popular in supporting pervasive applications, due to the demand for highly responsive and intuitive human control interfaces beyond the traditional keyboard and mouse. However, the heterogeneous nature of novel multimodal input devices and the tight coupling between input devices and applications complicate their deployment, rendering their dynamic integration to the intended applications rather difficult. i∗Chameleon exploits device abstraction in a web services-based framework to alleviate these problems. Developers can dynamically register new devices with the i∗Chameleon framework. They can also map specific device inputs to keyboard and mouse events efficiently. A number of input modalities such as tangible devices, speech, and finger gestures have been implemented to validate the feasibility of the i∗Chameleon framework in supporting multimodal input for pervasive applications.

In this paper, we propose a novel approach, Cyclic Signature (CS) clustering, to analyze spatial-temporal pattern. CS clustering is based on the calendar regularities of events to analyze spatial-temporal patterns. An experiment, based on a set of reported crime data for a district in Hong Kong, was performed to compare CS clustering against traditional clustering approaches. The results show that CS clustering can provide information which differs greatly from traditional clustering approaches. In addition, the groups created by CS clustering have higher intra-cluster similarities and lower inter-cluster similarities than traditional clustering approaches.

Controlling a team of robots is much more challenging than controlling an individual robot. Users desire high level commands that shield the users from too much detail, and yet still afford the desired precision of control. This project aims to design a method through which a team of robots can be controlled as easily and precisely as an individual robot. A simple language in the form of a set of finger gestures allows the user to give general motion commands to the team of robots. The gestures are supplemented by fine controls such as speed through tangible input gadgets. The gesture-based language has been implemented in a prototype user interface on a multi-touch screen. A number of test applications demonstrate the validity of the design.

This paper presents a design for a platform of collaborative robots and electronic devices. The platform design is based on the similarity between the type, functionality and characteristics of the robot or device. We categorize the participating devices by functionality rather than by architecture, therefore making it easy to support new robots or devices with similar functions but different architectures. This approach also allows users to develop, implement and port applications quickly and easily. We demonstrate the efficacy and correctness of our platform through a variety of robotic applications ranging from research to teaching.

Advances in textile-friendly electronics devices, smart materials, sophisticated interfaces, and intelligent software have combined to make intelligent garments increasingly practical. This paper introduces Co-eTex, a framework designed to support collaborative development of intelligent garments. It is based on a newly-invented construction platform for eTextiles and Wearable Computing, which was designed to be robust, reliable, easy to construct and to program. It also includes a hybrid graphical-textual programming tool designed for novice programmers to program intelligent behaviour for intelligent garments. We describe the use of Co-eTex in rapidly developing prototypes for a variety of designs of intelligent garments. Based on our experiences in using the Co-eTex, we believe it is a possible direction leading to the development of mass customization or adaptive customization of intelligent garments.

Airfreight forwarding is becoming more important with the increase of international trade. For an international company, sometimes, it would have several branches associating with different airports in the same region. When a branch is overloaded, nearby branches of different airports can offer help and share the workload. A framework of agents has been proposed to support collaborative cargo packing for multi-sites in the same region. Besides defining a set of XML message exchange standards in the framework, a core component is the packing module which is supported by RFID tags with information encoded. This paper discusses how to apply First Fit Decreasing and Best Fit Decreasing for packing airfreight cargoes and their preliminary performance.

Spatial analysis plays a key role in crime prevention. Traditional approaches such as clustering can find static patterns but do not consider the change of spatial patterns over time. In this paper, we introduce a new analysis framework, dynamic pattern analysis framework (DPA framework) focusing on two types of related dynamic patterns: the displacement or diffusion of spatial patterns over time and the similarity between spatial patterns of different periods. The new framework aims to support cooperative crime prevention in a district of Hong Kong.

Similar web pages of web sites on the World Wide Web are usually encoded from an underlying structured source, and generated dynamically from a pre-defined template, such as books' information pages in Amazon.com. By given a set of web pages from a common website, it is possible to extract the template by analyzing common patterns between the web pages. In our work, we developed the CF-EXALG (Collaborative Finer-EXALG), based on EXALG, to decompose web pages and finding their common structures. In our system, templates that are used to generate web pages can be discovered automatically and stored in XML format. Hence, data encoded in web pages can be easily extracted and the template can be stored for future manipulation. In our preliminary experiments, CF-EXALG has shown to be more accurate and efficient when compared with other similar systems.

XML is becoming a common way of storing data. The elements and their arrangement in the document’s hierarchy not only describe the document structure but also imply the data’s semantic meaning, and hence provide valuable information to develop tools for manipulating XML documents. In this paper, we pursue a data mining approach to the problem of XML document clustering. We introduce a novel XML structural representation called common XPath (CXP), which encodes the frequently occurring elements with the hierarchical information, and propose to take the CXPs mined to form the feature vectors for XML document clustering. In other words, data mining acts as a feature extractor in the clustering process. Based on this idea, we devise a path-based XML document clustering algorithm called PBClustering which groups the documents according to their CXPs, i.e. their frequent structures. Encouraging simulation results are observed and reported.

Reports our work in developing four new metrics to guide the development of computer-assisted learning (CAL) packages and to measure the the educational suitability of a given package. Our first metric, "multi-level compactness", is a summary index of the compactness at different hierarchies. The metric measures if teaching units are organized hierarchically while maintaining the local compactness within each teaching unit. The second metric, "heterogenity", is then developed to measure how different types of documents are clustered together in a CAL package. However, this metric does not reflect the structure of the different documents. This is remedied by the third metric, "guidance"; it represents how the exercises/test documents are placed in the traversal paths of the content and reference documents. The first three metrics are focused on the structure of the education materials. The forth metric, "crossing", is proposed to measure the help level amongst the materials. In order to verify the correctness of the first three metrics, a Java program has been developed. It provides an interface to calculate the values of the metrics for a given hyperlink of a WWW CAL package.

Real-time collaborative design of complex objects across great distances has been made much easier through the World Wide Web. A number of issues, however, need to be resolved. These include the sharing and exchange of information on the designed objects, the coordination and control of the collaborative design activities, and the management of the different design versions. This paper explores these three issues and describes a prototype system that resolves the issues in the context of the collaborative design of complex geometrical objects.