Human behavior recognition with generic exponential family duration modeling in the hidden semi-Markov model

The ability to learn and recognize human activities of daily living (ADLs) is important in building pervasive and smart environments. In this paper, we tackle this problem using the hidden semi-Markov model. We discuss the state-of-the-art duration modeling choices and then address a large class of exponential family distributions to model state durations. Inference and learning are efficiently addressed by providing a graphical representation for the model in terms of a dynamic Bayesian network (DBN). We investigate both discrete and continuous distributions from the exponential family (Poisson and Inverse Gaussian respectively) for the problem of learning and recognizing ADLs. A full comparison between the exponential family duration models and other existing models including the traditional multinomial and the new Coxian are also presented. Our work thus completes a thorough investigation into the aspect of duration modeling and its application to human activities recognition in a real-world smart home surveillance scenario.

The hidden permutation model and location-based activity recognition

Permutation modeling is challenging because of the combinatorial nature of the problem. However, such modeling is often required in many real-world applications, including activity recognition where subactivities are often permuted and partially ordered. This paper introduces a novel Hidden Permutation Model (HPM) that can learn the partial ordering constraints in permuted state sequences. The HPM is parameterized as an exponential family distribution and is flexible so that it can encode constraints via different feature functions. A chain-flipping Metropolis-Hastings Markov chain Monte Carlo (MCMC) is employed for inference to overcome the O(n!) complexity. Gradient-based maximum likelihood parameter learning is presented for two cases when the permutation is known and when it is hidden. The HPM is evaluated using both simulated and real data from a location-based activity recognition domain. Experimental results indicate that the HPM performs far better than other baseline models, including the naive Bayes classifier, the HMM classifier, and Kirshner's multinomial permutation model. Our presented HPM is generic and can potentially be utilized in any problem where the modeling of permuted states from noisy data is needed.

Multi-owner low-cost housing management in Malaysia : effects of owner-occupant characteristics and occupancy rates

Purpose – Homeownership is considered both economically and socially beneficial for homeowners. However, in the collective living arrangement, reaching a consensus with regard to the residential environment is difficult. The purpose of this paper is to identify factors that can reduce the conflict among the stakeholders in multi-owner low-cost housing in Malaysia.
Design/methodology/approach – This study tested three hypotheses examining whether the demographic and socio-economic characteristics of owner-occupants and occupancy rates affect owner-occupants' satisfaction with stakeholders' relationships. Data were collected through questionnaires from owner-occupants of multi-owner low-cost settlements in Selangor state. Data on housing characteristics were collected from chairpersons of the respective owners' organisations. The data were treated as parametric, and analysis of variance was conducted.
Findings – Four factors – number of children in the family, duration of residency, participation in social activities and participation in meetings – were found to affect owners-occupants' satisfaction with the stakeholders' relationships. The significant effect of occupancy rates was also indicated.
Practical implications – The Management Corporations (MCs) should encourage social relationships among residents. To avoid conflict, the costs and benefits of participation must be balanced. Policy makers should take two key aspects seriously: owner-managed strategy practices by the MCs and high rates of tenant-residents. A mechanism should be identified for assisting the MCs in housing management and for protecting the benefits of homeownership for owner-occupants.
Originality/value – Past studies on low-income household settlements examined public housing or low-income homeowners of single detached dwellings. This study adds to the existing body of knowledge by examining low-income homeowners in multi-owner low-cost settlements.

A 'respectful relationships' approach : could it be the answer to preventing gender-based violence?

The probability model of peer-to-peer botnet propagation

Active Peer-to-Peer worms are great threat to the network security since they can propagate in automated ways and flood the Internet within a very short duration. Modeling a propagation process can help us to devise effective strategies against a worm's spread. This paper presents a study on modeling a worm's propagation probability in a P2P overlay network and proposes an optimized patch strategy for defenders. Firstly, we present a probability matrix model to construct the propagation of P2P worms. Our model involves three indispensible aspects for propagation: infected state, vulnerability distribution and patch strategy. Based on a fully connected graph, our comprehensive model is highly suited for real world cases like Code Red II. Finally, by inspecting the propagation procedure, we propose four basic tactics for defense of P2P botnets. The rationale is exposed by our simulated experiments and the results show these tactics are of effective and have considerable worth in being applied in real-world networks.

Topic transition detection using hierarchical hidden Markov and semi-Markov models

In this paper we introduce a probabilistic framework to exploit hierarchy, structure sharing and duration information for topic transition detection in videos. Our probabilistic detection framework is a combination of a shot classification step and a detection phase using hierarchical probabilistic models. We consider two models in this paper: the extended Hierarchical Hidden Markov Model (HHMM) and the Coxian Switching Hidden semi-Markov Model (S-HSMM) because they allow the natural decomposition of semantics in videos, including shared structures, to be modeled directly, and thus enabling efficient inference and reducing the sample complexity in learning. Additionally, the S-HSMM allows the duration information to be incorporated, consequently the modeling of long-term dependencies in videos is enriched through both hierarchical and duration modeling. Furthermore, the use of the Coxian distribution in the S-HSMM makes it tractable to deal with long sequences in video. Our experimentation of the proposed framework on twelve educational and training videos shows that both models outperform the baseline cases (flat HMM and HSMM) and performances reported in earlier work in topic detection. The superior performance of the S-HSMM over the HHMM verifies our belief that duration information is an important factor in video content modeling.

Magával ragadó interfészek Art Education tanítás és tanulás vegyes valóság virtuális és valós környezetben

How selected next generation technologies support collaborative participation between higher education students and educators within a virtual socially networked e-learning landscape and encourage the interaction of communities of learners in multiple modes, ranging from text and images accessed within the Deakin Studies Online learning management system to a constructed virtual world in which the user’s creative imagination transports them to the “other side” of their computer screens is discussed in this paper. These constructed environments enable multiple simultaneous participants to access graphically built 3D environments, interact with digital artifacts and various functional tools and represent themselves through avatars, to communicate with other participants and engage in collaborative art learning.

The best of all worlds : immersive interfaces for art education in virtual and real world teaching and learning environments

A collaborative constructivist model of e-learning enabled second year undergraduate students and art educators to establish a community of learners within an augmented immersive learning environment. Artistic practice and work based learning was enhanced through the creation of digital artifacts to support shared knowledge building using authentic learning tasks and social networking.

Immersive interfaces for art education teaching and learning in virtual and real world learning environments

The business of sexuality education : changing practice : the need for professional development

This paper reports on data collected from practicing sexuality education teachers following their participation in professional learning intervention in sexuality education. It explores whether the provision of effective professional development and classroom resources enables teachers to effectively address the sensitive issues of sexual diversity, gender and power.

Pauline Kael and the Western genre as critical displacement of self & nation: metaphorics and affects of 'taste' in American film criticism

Pauline Kael (1919–2001) is one of the most influential American film critics of the second half of the twentieth century. Many people are writing on her presently, with at least half an eye to her future cultural, political and historical importance. Certainly the full impact of Kael’s work, both within and beyond the borders of cinema (however defined), has not yet been established. This article unpacks the mechanisms and operations of ‘taste’ in Kael’s writings by using two notions drawn from Roland Barthes’ observations about another key figure of current cultural critique: Julia Kristeva. The comparison of Kael with Kristeva is not dwelt upon; instead, the article focuses on how Kael used the concepts of ‘taste’ and ‘dis-taste’ to draw her readership into a field of what might be termed ‘permanent dissent’. This article concludes by sketching out why Jewish-American Kael’s taste might endure, through the dual transition she occupies from a Cold War to a post-Cold War period, and from an era when cinema was the supreme, undisputed, screen artform, to the rise of the myriad screen technologies of the networked, Internet age.

Comparing personality type profiles of primary teachers in the Catholic and state school systems

A new transfer learning framework with application to model-agnostic multi-task learning

Learning from small number of examples is a challenging problem in machine learning. An effective way to improve the performance is through exploiting knowledge from other related tasks. Multi-task learning (MTL) is one such useful paradigm that aims to improve the performance through jointly modeling multiple related tasks. Although there exist numerous classification or regression models in machine learning literature, most of the MTL models are built around ridge or logistic regression. There exist some limited works, which propose multi-task extension of techniques such as support vector machine, Gaussian processes. However, all these MTL models are tied to specific classification or regression algorithms and there is no single MTL algorithm that can be used at a meta level for any given learning algorithm. Addressing this problem, we propose a generic, model-agnostic joint modeling framework that can take any classification or regression algorithm of a practitioner’s choice (standard or custom-built) and build its MTL variant. The key observation that drives our framework is that due to small number of examples, the estimates of task parameters are usually poor, and we show that this leads to an under-estimation of task relatedness between any two tasks with high probability. We derive an algorithm that brings the tasks closer to their true relatedness by improving the estimates of task parameters. This is achieved by appropriate sharing of data across tasks. We provide the detail theoretical underpinning of the algorithm. Through our experiments with both synthetic and real datasets, we demonstrate that the multi-task variants of several classifiers/regressors (logistic regression, support vector machine, K-nearest neighbor, Random Forest, ridge regression, support vector regression) convincingly outperform their single-task counterparts. We also show that the proposed model performs comparable or better than many state-of-the-art MTL and transfer learning baselines.

Finite element modeling of a generic stemless hip implant design in comparison with conventional hip implants

In this study, the finite element modeling and comparison of the stress and strain analyses were carried out for three different structures that are intact bone, stemless implant and stemmed one. Currently proposed stemless design studied here is the generic concept of stemless implant. This generic stemless implant reconstruction was numerically compared to the conventional stemmed implant and also to the intact bone as control solution. Two loading conditions were applied to the most proximal part of the models, while the most distal part was fixed for all degrees of freedom. The models were divided into two regions and studied along two paths of medial and lateral aspect. The results of this study showed that the stemless implant had less deviation from the control solution of the bone in all regions and in both loading conditions, comparing to the large deviation of the stemmed implant from the intact bone. However, it was shown that the fixation of this type of implant and its effect on sub-trochanter region must be carefully considered for designing the final product of any specific design of stemless implant.