Learning latent activities from social signals with hierarchical dirichlet processes

Understanding human activities is an important research topic, most noticeably in assisted-living and healthcare monitoring environments. Beyond simple forms of activity (e.g., an RFID event of entering a building), learning latent activities that are more semantically interpretable, such as sitting at a desk, meeting with people, or gathering with friends, remains a challenging problem. Supervised learning has been the typical modeling choice in the past. However, this requires labeled training data, is unable to predict never-seen-before activity, and fails to adapt to the continuing growth of data over time. In this chapter, we explore the use of a Bayesian nonparametric method, in particular the hierarchical Dirichlet process, to infer latent activities from sensor data acquired in a pervasive setting. Our framework is unsupervised, requires no labeled data, and is able to discover new activities as data grows. We present experiments on extracting movement and interaction activities from sociometric badge signals and show how to use them for detecting of subcommunities. Using the popular Reality Mining dataset, we further demonstrate the extraction of colocation activities and use them to automatically infer the structure of social subgroups. © 2014 Elsevier Inc. All rights reserved.

Least squares learning and the US Treasury bill rate

Understanding how agents formulate their expectations about Fed behavior is important for market participants because they can potentially use this information to make more accurate estimates of stock and bond prices. Although it is commonly assumed that agents learn over time, there is scant empirical evidence in support of this assumption. Thus, in this paper we test if the forecast of the three month T-bill rate in the Survey of Professional Forecasters (SPF) is consistent with least squares learning when there are discrete shifts in monetary policy. We first derive the mean, variance and autocovariances of the forecast errors from a recursive least squares learning algorithm when there are breaks in the structure of the model. We then apply the Bai and Perron (1998) test for structural change to a forecasting model for the three month T-bill rate in order to identify changes in monetary policy. Having identified the policy regimes, we then estimate the implied biases in the interest rate forecasts within each regime. We find that when the forecast errors from the SPF are corrected for the biases due to shifts in policy, the forecasts are consistent with least squares learning. © 2014 Elsevier B.V.

Confronting globalisation: learning from intercontinental collaboration

Higher education institutions are responding to globalisation in various ways. This study describes and analyses challenges encountered in a recent case of global collaboration between four universities on different continents in developing a web‐based master’s program. The key issue was how to develop programs in a way that is fair for the different countries involved. The focus of the paper is on tensions between local and national contexts, rules and resources and the creation of a common global program. ‘Agency’, ‘structure’ and ‘frame factor’ are used as analytical concepts to help understand the dynamics of the collaboration and the character of the program.

Corrigendum to “Least squares learning and the US treasury bill rate” [Econ. Syst. 38 (2014) 194–204]

Understanding how agents formulate their expectations about Fed behavior is important for market participants because they can potentially use this information to make more accurate estimates of stock and bond prices. Although it is commonly assumed that agents learn over time, there is scant empirical evidence in support of this assumption. Thus, in this paper we test if the forecast of the three month T-bill rate in the Survey of Professional Forecasters (SPF) is consistent with least squares learning when there are discrete shifts in monetary policy. We first derive the mean, variance and autocovariances of the forecast errors from a recursive least squares learning algorithm when there are breaks in the structure of the model. We then apply the Bai and Perron (1998) test for structural change to a forecasting model for the three month T-bill rate in order to identify changes in monetary policy. Having identified the policy regimes, we then estimate the implied biases in the interest rate forecasts within each regime. We find that when the forecast errors from the SPF are corrected for the biases due to shifts in policy, the forecasts are consistent with least squares learning.

A systematic design of interval type-2 fuzzy logic system using extreme learning machine for electricity load demand forecasting

This paper presents a novel design of interval type-2 fuzzy logic systems (IT2FLS) by utilizing the theory of extreme learning machine (ELM) for electricity load demand forecasting. ELM has become a popular learning algorithm for single hidden layer feed-forward neural networks (SLFN). From the functional equivalence between the SLFN and fuzzy inference system, a hybrid of fuzzy-ELM has gained attention of the researchers. This paper extends the concept of fuzzy-ELM to an IT2FLS based on ELM (IT2FELM). In the proposed design the antecedent membership function parameters of the IT2FLS are generated randomly, whereas the consequent part parameters are determined analytically by the Moore-Penrose pseudo inverse. The ELM strategy ensures fast learning of the IT2FLS as well as optimality of the parameters. Effectiveness of the proposed design of IT2FLS is demonstrated with the application of forecasting nonlinear and chaotic data sets. Nonlinear data of electricity load from the Australian National Electricity Market for the Victoria region and from the Ontario Electricity Market are considered here. The proposed model is also applied to forecast Mackey-glass chaotic time series data. Comparative analysis of the proposed model is conducted with some traditional models such as neural networks (NN) and adaptive neuro fuzzy inference system (ANFIS). In order to verify the structure of the proposed design of IT2FLS an alternate design of IT2FLS based on Kalman filter (KF) is also utilized for the comparison purposes.

Less is more: development and evaluation of an interactive e-atlas to support anatomy learning

An Interactive electronic Atlas (IeA) was developed to assist first-year nursing students with interpretation of laboratory-based prosected cadaveric material. It was designed, using pedagogically sound principles, as a student-centered resource accessible to students from a wide range of learning backgrounds. It consisted of a highly simplified interactive interface limited to essential anatomical structures and was intended for use in a blended learning situation. The IeA's nine modules mirrored the body systems covered in a Nursing Biosciences course, with each module comprising a maximum of 10 pages using the same template: an image displaying a cadaveric specimen and, in most cases, a corresponding anatomical model with navigation panes (menus) on one side. Cursor movement over the image or clicking the menu highlighted the structure with a transparent overlay and revealed a succinct functional description. The atlas was complemented by a multiple-choice database of nearly 1,000 questions using IeA images. Students' perceptions of usability and utility were measured by survey (n = 115; 57% of the class) revealing mean access of 2.3 times per week during the 12-week semester and a median time of three hours of use. Ratings for usability and utility were high, with means ranging between 4.24 and 4.54 (five-point Likert scale; 5 = strongly agree). Written responses told a similar story for both usability and utility. The role of providing basic computer-assisted learning support for a large first-year class is discussed in the context of current research into student-centered resources and blended learning in human anatomy.

How online learning in an engineering-physics course helped us flip the classroom

Distance education has developed in the past 25 years or so as a way of supplying education to people who would not have access to local college education facilities. This includes students who live in remote regions, students who lack mobility, and students with full-time jobs. More recently this has been renamed to "online learning". Deakin University in Australia has been teaching freshman engineering physics simultaneously to on-campus and online students since the late1990's. The course is part of an online Bachelor of Engineering major that is accredited by the Institution of Engineers Australia.* In this way Deakin answers the call to provide engineering education "anywhere, anytime."**The course has developed and improved with the available educational technology. Starting with printed study guides, a textbook, CD-ROMS, and snail-mail, and telephone/email correspondence with students, the course has seen the rise of websites, online course notes, discussion boards, streamed video lectures, web-conferencing classes and lab sessions, and online submission of student work. Most recently the on-campus version of the course has shifted from a traditional lecture/tutorial/lab format to a flipped-classroom format. The use of lectures has been reduced while the use of tutorials and practical exercises has increased. Primary learning is now accomplished by watching videos prepared by the lecturer and studying the textbook.Offering this course for several years by distance education made this process considerably easier. Most of the educational "infrastructure" was already in place, and the course's delivery to a non-classroom cohort was already established. Thus many elements of the new structure did not have to be produced from scratch. Improvements to the course website and all the course material has benefited all students, both online and on-campus.The new course structure was delivered for the first time in 2014, has run for two semesters, and will continue in 2015. Student learning and performance is being measured by assignment and exam marks for both on-campus and off-campus students. Students are also surveyed to gauge how well they received the new innovations, especially the video presentations on the lab experiments. It was found that student performance in the new structure was no worse than that in the older structure (average on-campus grades increased 10%), and students in general welcomed the changes. Similar transitions are being implemented in other courses in Deakin's engineering degree program.This presentation will show how physics is taught to online students, outline the changes made to support flipping the on-campus classroom, and how that process benefited the off-campus cohort.

Engineering education anywhere, anytime: from distance education to blended learning at Deakin University Australia

In 2005 the Sloan Consortium called for engineering education to be available "anywhere, anytime."* Increasing numbers of engineering departments are interesting in offering their programs by means of online learning. These schools grapple with several difficulties and issues associated with wholly online learning: course structure, communication with students, delivery of course material, delivery of exams, accreditation, equity between on-campus and off-campusstudents, and especially the delivery of practical training. Deakin University faced these same challenges when it commenced teaching undergraduate engineering via distance education in the early 1990's. It now offers a fully accredited Bachelor of Engineering degree in both on-campus and off-campus modes, with majors that include civil,mechanical, electrical/electronics, and mechatronics/robotics.This presentation describes Deakin's unique off-campus delivery, students, curricula, approaches to practical work, and solutions to the problems mentioned above. Attendees will experience how Deakin Engineering delivers course materials, communicates with off-campus students, runs off-campus classes, and even delivers lab experience to students living thousands of miles away from the home campus. On display will be experimental lab kits, video presentations, student projects, and online broadcasts of freshman lab experiments. Participants will have the opportunity to see some of these resources hands-on. I will also discuss recent innovations in off-campus delivery ofcourses, including how flipping the classroom has led to blended learning with the on-campus students.Many universities have placed engineering distance education into the too-hard basket. Deakin Engineering demonstrates that it is possible to deliver a full undergraduate degree by means of distance education and online learning, and modern technology makes the job easier than everbefore. The benefits to the professor are many, not the least of which is helping a student living in a remote area or with a full-time job become fully trained and qualified in engineering.

Identifying inorganic material affinity classes for peptide sequences based on context learning

There is a growing interest in identifying inorganic material affinity classes for peptide sequences due to the development of bionanotechnology and its wide applications. In particular, a selective model capable of learning cross-material affinity patterns can help us design peptide sequences with desired binding selectivity for one inorganic material over another. However, as a newly emerging topic, there are several distinct challenges of it that limit the performance of many existing peptide sequence classification algorithms. In this paper, we propose a novel framework to identify affinity classes for peptide sequences across inorganic materials. After enlarging our dataset by simulating peptide sequences, we use a context learning based method to obtain the vector representation of each amino acid and each peptide sequence. By analyzing the structure and affinity class of each peptide sequence, we are able to capture the semantics of amino acids and peptide sequences in a vector space. At the last step we train our classifier based on these vector features and the heuristic rules. The construction of our models gives us the potential to overcome the challenges of this task and the empirical results show the effectiveness of our models.

A fundamental rethink of practice learning in social work education

A common response to the need to place increasing numbers of social work students in field education or practice learning placements has been to broaden the range of organisations in which placements are sought. While this strategy has provided many beneficial learning opportunities for students, it has not been sufficient in tackling ongoing difficulties in locating work-integrated learning opportunities for social work students. We argue that new approaches to finding placement opportunities will require a fundamental rethink as to how student placements are understood. This paper introduces an innovative project which started with a consideration of learning opportunities and built a structure to facilitate these, rather than rely on organisational availability to host students on placements.

Learning multi-faceted activities from heterogeneous data with the product space hierarchical Dirichlet processes

Hierarchical Dirichlet processes (HDP) was originally designed and experimented for a single data channel. In this paper we enhanced its ability to model heterogeneous data using a richer structure for the base measure being a product-space. The enhanced model, called Product Space HDP (PS-HDP), can (1) simultaneously model heterogeneous data from multiple sources in a Bayesian nonparametric framework and (2) discover multilevel latent structures from data to result in different types of topics/latent structures that can be explained jointly. We experimented with the MDC dataset, a large and real-world data collected from mobile phones. Our goal was to discover identity–location– time (a.k.a who-where-when) patterns at different levels (globally for all groups and locally for each group). We provided analysis on the activities and patterns learned from our model, visualized, compared and contrasted with the ground-truth to demonstrate the merit of the proposed framework. We further quantitatively evaluated and reported its performance using standard metrics including F1-score, NMI, RI, and purity. We also compared the performance of the PS-HDP model with those of popular existing clustering methods (including K-Means, NNMF, GMM, DP-Means, and AP). Lastly, we demonstrate the ability of the model in learning activities with missing data, a common problem encountered in pervasive and ubiquitous computing applications.