Supervised latent dirichlet allocation models for efficient activity representation

Local spatio-temporal features with a Bag-of-visual words model is a popular approach used in human action recognition. Bag-of-features methods suffer from several challenges such as extracting appropriate appearance and motion features from videos, converting extracted features appropriate for classification and designing a suitable classification framework. In this paper we address the problem of efficiently representing the extracted features for classification to improve the overall performance. We introduce two generative supervised topic models, maximum entropy discrimination LDA (MedLDA) and class- specific simplex LDA (css-LDA), to encode the raw features suitable for discriminative SVM based classification. Unsupervised LDA models disconnect topic discovery from the classification task, hence yield poor results compared to the baseline Bag-of-words framework. On the other hand supervised LDA techniques learn the topic structure by considering the class labels and improve the recognition accuracy significantly. MedLDA maximizes likelihood and within class margins using max-margin techniques and yields a sparse highly discriminative topic structure; while in css-LDA separate class specific topics are learned instead of common set of topics across the entire dataset. In our representation first topics are learned and then each video is represented as a topic proportion vector, i.e. it can be comparable to a histogram of topics. Finally SVM classification is done on the learned topic proportion vector. We demonstrate the efficiency of the above two representation techniques through the experiments carried out in two popular datasets. Experimental results demonstrate significantly improved performance compared to the baseline Bag-of-features framework which uses kmeans to construct histogram of words from the feature vectors.

Murine, but not human, ephrin-B2 can be efficiently cleaved by the serine protease kallikrein-4: Implications for xenograft models of human prostate cancer

Background Ephrin-B2 is the sole physiologically-relevant ligand of the receptor tyrosine kinase EphB4, which is over-expressed in many epithelial cancers, including 66% of prostate cancers, and contributes to cancer cell survival, invasion and migration. Crucially, however, the cancer-promoting EphB4 signalling pathways are independent of interaction with its ligand ephrin-B2, as activation of ligand-dependent signalling causes tumour suppression. Ephrin-B2, however, is often found on the surface of endothelial cells of the tumour vasculature, where it can regulate angiogenesis to support tumour growth. Proteolytic cleavage of endothelial cell ephrin-B2 has previously been suggested as one mechanism whereby the interaction between tumour cell-expressed EphB4 and endothelial cell ephrin-B2 is regulated to support both cancer promotion and angiogenesis. Methods An in silico approach was used to search accessible surfaces of 3D protein models for cleavage sites for the key prostate cancer serine protease, KLK4, and this identified murine ephrin-B2 as a potential KLK4 substrate. Mouse ephrin-B2 was then confirmed as a KLK4 substrate by in vitro incubation of recombinant mouse ephrin-B2 with active recombinant human KLK4. Cleavage products were visualised by SDS-PAGE, silver staining and Western blot and confirmed by N-terminal sequencing. Results At low molar ratios, KLK4 cleaved murine ephrin-B2 but other prostate-specific KLK family members (KLK2 and KLK3/PSA) were less efficient, suggesting cleavage was KLK4-selective. The primary KLK4 cleavage site in murine ephrin-B2 was verified and shown to correspond to one of the in silico predicted sites between extracellular domain residues arginine 178 and asparagine 179. Surprisingly, the highly homologous human ephrin-B2 was poorly cleaved by KLK4 at these low molar ratios, likely due to the 3 amino acid differences at this primary cleavage site. Conclusion These data suggest that in in vivo mouse xenograft models, endogenous mouse ephrin-B2, but not human tumour ephrin-B2, may be a downstream target of cancer cell secreted human KLK4. This is a critical consideration when interpreting data from murine explants of human EphB4+/KLK4+ cancer cells, such as prostate cancer cells, where differential effects may be seen in mouse models as opposed to human clinical situations.

Active learning: A step towards automating medical concept extraction

Objective This paper presents an automatic active learning-based system for the extraction of medical concepts from clinical free-text reports. Specifically, (1) the contribution of active learning in reducing the annotation effort, and (2) the robustness of incremental active learning framework across different selection criteria and datasets is determined. Materials and methods The comparative performance of an active learning framework and a fully supervised approach were investigated to study how active learning reduces the annotation effort while achieving the same effectiveness as a supervised approach. Conditional Random Fields as the supervised method, and least confidence and information density as two selection criteria for active learning framework were used. The effect of incremental learning vs. standard learning on the robustness of the models within the active learning framework with different selection criteria was also investigated. Two clinical datasets were used for evaluation: the i2b2/VA 2010 NLP challenge and the ShARe/CLEF 2013 eHealth Evaluation Lab. Results The annotation effort saved by active learning to achieve the same effectiveness as supervised learning is up to 77%, 57%, and 46% of the total number of sequences, tokens, and concepts, respectively. Compared to the Random sampling baseline, the saving is at least doubled. Discussion Incremental active learning guarantees robustness across all selection criteria and datasets. The reduction of annotation effort is always above random sampling and longest sequence baselines. Conclusion Incremental active learning is a promising approach for building effective and robust medical concept extraction models, while significantly reducing the burden of manual annotation.

External knowledge and query strategies in active learning: A study in clinical information extraction

This paper presents a new active learning query strategy for information extraction, called Domain Knowledge Informativeness (DKI). Active learning is often used to reduce the amount of annotation effort required to obtain training data for machine learning algorithms. A key component of an active learning approach is the query strategy, which is used to iteratively select samples for annotation. Knowledge resources have been used in information extraction as a means to derive additional features for sample representation. DKI is, however, the first query strategy that exploits such resources to inform sample selection. To evaluate the merits of DKI, in particular with respect to the reduction in annotation effort that the new query strategy allows to achieve, we conduct a comprehensive empirical comparison of active learning query strategies for information extraction within the clinical domain. The clinical domain was chosen for this work because of the availability of extensive structured knowledge resources which have often been exploited for feature generation. In addition, the clinical domain offers a compelling use case for active learning because of the necessary high costs and hurdles associated with obtaining annotations in this domain. Our experimental findings demonstrated that 1) amongst existing query strategies, the ones based on the classification model’s confidence are a better choice for clinical data as they perform equally well with a much lighter computational load, and 2) significant reductions in annotation effort are achievable by exploiting knowledge resources within active learning query strategies, with up to 14% less tokens and concepts to manually annotate than with state-of-the-art query strategies.

Enrichment of circulating interleukin-17-secreting interleukin-23 receptor-positive γ/δ T cells in patients with active ankylosing spondylitis

Objective Ankylosing spondylitis (AS) is a common inflammatory arthritis affecting primarily the axial skeleton. IL23R is genetically associated with AS. This study was undertaken to investigate and characterize the role of interleukin-23 (IL-23) signaling in AS pathogenesis. Methods The study population consisted of patients with active AS (n = 17), patients with psoriatic arthritis (n = 8), patients with rheumatoid arthritis, (n = 9), and healthy subjects (n = 20). IL-23 receptor (IL-23R) expression in T cells was determined in each subject group, and expression levels were compared. Results The proportion of IL-23R-expressing T cells in the periphery was 2-fold higher in AS patients than in healthy controls, specifically driven by a 3-fold increase in IL-23R-positive γ/δ T cells in AS patients. The proportions of CD4+ and CD8+ cells that were positive for IL-17 were unchanged. This increased IL-23R expression on γ/δ T cells was also associated with enhanced IL-17 secretion, with no observable IL-17 production from IL-23R-negative γ/δ T cells in AS patients. Furthermore, γ/δ T cells from AS patients were heavily skewed toward IL-17 production in response to stimulation with IL-23 and/or anti-CD3/CD28. Conclusion Recently, mouse models have shown IL-17-secreting γ/δ T cells to be pathogenic in infection and autoimmunity. Our data provide the first description of a potentially pathogenic role of these cells in a human autoimmune disease. Since IL-23 is a maturation and growth factor for IL-17-producing cells, increased IL-23R expression may regulate the function of this putative pathogenic γ/δ T cell population.

Use of semi-active dampers in seismic mitigation of building structures

This research provides information for providing the required seismic mitigation in building structures through the use of semi active and passive dampers. The Magneto-Rheological (MR) semi-active damper model was developed using control algorithms and integrated into seismically excited structures as a time domain function. Linear and nonlinear structure models are evaluated in real time scenarios. Research information can be used for the design and construction of earthquake safe buildings with optimally employed MR dampers and MR-passive damper combinations.