Human action categorization using conditional random field

Automatic human action recognition has been a challenging issue in the field of machine vision. Some high-level features such as SIFT, although with promising performance for action recognition, are computationally complex to some extent. To deal with this problem, we construct the features based on the Distance Transform of body contours, which is relatively simple and computationally efficient, to represent human action in the video. After extracting the features from videos, we adopt the Conditional Random Field for modeling the temporal action sequences. The proposed method is tested with an available standard dataset. We also testify the robustness of our method on various realistic conditions, such as body occlusion or intersection.

Hierarchical semi-markov conditional random fields for recursive sequential data

Inspired by the hierarchical hidden Markov models (HHMM), we present the hierarchical semi-Markov conditional random field (HSCRF), a generalisation of embedded undirected Markov chains to model complex hierarchical, nested Markov processes. It is parameterised in a discriminative framework and has polynomial time algorithms for learning and inference. Importantly, we develop efficient algorithms for learning and constrained inference in a partially-supervised setting, which is important issue in practice where labels can only be obtained sparsely. We demonstrate the HSCRF in two applications: (i) recognising human activities of daily living (ADLs) from indoor surveillance cameras, and (ii) noun-phrase chunking. We show that the HSCRF is capable of learning rich hierarchical models with reasonable accuracy in both fully and partially observed data cases.

Activating teacher critical moments of learning through reflection

Mathematical modelling is one of the current focuses in the Singapore Mathematics Curriculum Framework. A multi-tiered teaching experiment using design research methodology was conducted to build teachers’ capacity in designing, facilitating, and evaluating learning during mathematical modelling tasks for Primary 5 students (aged 10-11). This paper illustrates the use of the retrospective analysis phase within design research cycles to activate a critical moment of teacher learning involving the interplay between questioning and listening during her first attempt at facilitating a mathematical modelling task. The teacher affirmed her deliberate focuses in the use of questions to (a) refine students’ models, (b) encourage articulation of student ideas in self-evaluation of the models, and (c) clarify and understand student reasoning. However, she also discovered the importance of interpretative listening in conjunction with questioning to promote more sophisticated mathematisation processes in model development. Implications from the use of the retrospective analysis phase on activating critical moments of learning during teacher education will be discussed.

Eureka moments in research : exploring abductive processes using four case examples

Abduction plays an important but often unacknowledged role in research – this regrettably leaving a large part of the research process hidden and unexamined (Levin-Rozalis 2000), particularly important innovative or creative components. This paper, firstly, introduces abduction and discusses some important concepts related to abduction and innovation. Secondly, it presents author’s own re-descriptions of previous research work. This new description seeks to describe perceived key “Eureka” moments in the research and thus make the creative components and abductive elements more visible. The paper demonstrates that much can be gained from opening a reflective space for the role of abduction in the research process.

Moments of disruption and the development of expatriate TESOL teachers

TESOL teachers, like mainstream teachers, often experience key incidents in their professional development. In expatriate TESOL, however, unfamiliar cultural and linguistic contexts may disrupt teachers' sense of both professional and personal identity. In this article, narratives constructed from interviews of teacher experiences document a selection of critical events and discuss their implications for professional development in TESOL. Teachers reported that deep reflection on their experiences led to a re-conceptualisation of their professional and cultural identities. The analysis of their reflections may have significant implications for TESOL work in the context of the global and the local. The narratives specifically explore teachers' initial experiences of readjustment, awareness of differences in expectations of their roles as teachers, the impact of being perceived as representatives of 'Western' culture, and their re-appraisal of the educational and cultural-linguistic values of 'home'.

Bone marrow chimeras and c-fms conditional ablation (mafia) mice reveal an essential role for resident myeloid cells in lipopolysaccharide/TLR4-induced corneal inflammation

The mammalian cornea contains an extensive network of resident macrophages and dendritic cells. To determine the role of these cells in LPS-induced corneal inflammation, TLR4−/− mice were sublethally irradiated and reconstituted with bone marrow cells from either enhanced GFP (eGFP)+/C57BL/6 or eGFP+/TLR4−/− mice. The corneal epithelium was abraded, LPS was added topically, and cellular infiltration to the corneal stroma and development of corneal haze were examined after 24 h. TLR4−/− mice reconstituted with C57BL/6, but not TLR4−/− bone marrow cells donor cells were found to cause infiltration of eGFP+ cells to the cornea, including neutrophils, and also increased corneal haze compared with saline-treated corneas. In a second experimental approach, corneas of transgenic macrophage Fas induced apoptosis (Mafia) mice were stimulated with LPS. These mice express eGFP and a suicide gene under control of the c-fms promoter, and systemic treatment with the FK506 dimerizer (AP20187) causes Fas-mediated apoptosis of monocytic cells. AP20187-treated mice had significantly fewer eGFP+ cells in the cornea than untreated mice. After stimulation with LPS neutrophil recruitment and development of corneal haze were impaired in AP20187-treated mice compared with untreated controls. Furthermore, LPS induced CXCL1/KC and IL-1α production within 4 h in corneas of untreated Mafia mice, which is before cellular infiltration; however, cytokine production was impaired after AP20187 treatment. Together, results from both experimental approaches demonstrate an essential role for resident corneal monocytic lineage cells (macrophages and dendritic cells) in development of corneal inflammation.

A computationally convenient unit root test with covariates, conditional heteroskedasticity and efficient detrending

When testing for a unit root in a time series, in spite of the well-known power problem of univariate tests, it is quite common to use only the information regarding the autoregressive behaviour contained in that series. In a series of influential papers, Elliott et al. (Efficient tests for an autoregressive unit root, Econometrica 64, 813–836, 1996), Hansen (Rethinking the univariate approach to unit root testing: using covariates to increase power, Econometric Theory 11, 1148–1171, 1995a) and Seo (Distribution theory for unit root tests with conditional heteroskedasticity, Journal of Econometrics 91, 113–144, 1999) showed that this practice can be rather costly and that the inclusion of the extraneous information contained in the near-integratedness of many economic variables, their heteroskedasticity and their correlation with other covariates can lead to substantial power gains. In this article, we show how these information sets can be combined into a single unit root test.

Non-nested tests of a GDP-augmented Fama–French model versus a conditional Fama–French model in the Australian stock market

We extend Vassalou (2003) by conditioning the Fama–French model with the same macroeconomic variables used to construct a GDP factor. The motivation for doing so is to ascertain whether the ability of the GDP-augmented model to explain equity returns is actually due to news about future GDP growth or whether it is due to the macroeconomic conditioning variables used to construct the GDP factor. We compare the performance of a GDP-enhanced Fama–French model with the conditional Fama–French model using non-nested testing techniques. We find that the GDP-augmented model considerably underperforms the conditional version of the model.