A slice sampler for restricted hierarchical beta process with applications to shared subspace learning

Hierarchical beta process has found interesting applications in recent years. In this paper we present a modified hierarchical beta process prior with applications to hierarchical modeling of multiple data sources. The novel use of the prior over a hierarchical factor model allows factors to be shared across different sources. We derive a slice sampler for this model, enabling tractable inference even when the likelihood and the prior over parameters are non-conjugate. This allows the application of the model in much wider contexts without restrictions. We present two different data generative models – a linear Gaussian-Gaussian model for real valued data and a linear Poisson-gamma model for count data. Encouraging transfer learning results are shown for two real world applications – text modeling and content based image retrieval.

Vitamin D and health in pregnancy, infants, children and adolescents in Australia and New Zealand: a position statement.

The recommended level for serum 25-hydroxyvitamin D (25(OH)D) in infants,  children,  dolescents and during pregnancy and lactation is ≥ 50 nmol/L. This level may need to be 10-20 nmol/L higher at the end of summer to maintain levels ≥ 50 nmol/L over winter and spring. • Sunlight is the most important source of vitamin D. The US recommended dietary allowance for vitamin D is 600 IU daily in children aged over 12 months and during pregnancy and lactation, assuming minimal sun exposure. • Risk factors for low vitamin D are: lack of skin exposure to sunlight, dark skin, southerly latitude, conditions affecting vitamin D metabolism and storage (including obesity) and, for infants, being born to a mother with low vitamin D and exclusive breastfeeding combined with at least one other risk factor. • Targeted measurement of 25(OH)D levels is recommended for infants, children and adolescents with at least one risk factor for low vitamin D and for pregnant women with at least one risk factor for low vitamin D at the first antenatal visit. • Vitamin D deficiency can be treated with daily low-dose vitamin D supplements, although barriers to adherence have been identified. High-dose intermittent vitamin D can be used in children and adolescents. Treatment should be paired with health education and advice about sensible sun exposure. Infants at risk of low vitamin D should be supplemented with 400 IU vitamin D₃ daily for at least the first year of life. • There is increasing evidence of an association between low vitamin D and a range of non-bone health outcomes, however there is a lack of data from robust randomised controlled trials of vitamin D supplementation.

Electrospun spinel LiNi0.5Mn1.5O4 hierarchical nanofibers as 5 v cathode materials for lithium-ion batteries

Spinel LiNi0.5Mn1.5O4 hierarchical nanofibers with diameters of 200–500 nm and lengths of up to several tens of micrometers were synthesized using low-cost starting materials by electrospinning combined with annealing. Well-separated nanofiber precursors impede the growth and agglomeration of Li-Ni0.5Mn1.5O4 particles. The hierarchical nanofibers were constructed from attached LiNi0.5Mn1.5O4 nanooctahedrons with sizes ranging from 200 to 400 nm. It is proven that these Li-Ni0.5Mn1.5O4 hierarchical nanofibers exhibit a favorable electrochemical performance. At a 0.5C (coulombic) rate, it shows an initial discharge capacity of 133 mAhg_1 with a capacity retention over 94% after 30 cycles. Even at 2, 5, 10, and 15C rates, it can still deliver a discharge capacity of 115, 100, 90, and 80 mAhg_1, respectively. Compared with self-aggregated nanooctahedrons synthesized using common sol–gel methods, the LiNi0.5Mn1.5O4 hierarchical nanofibers exhibit a much higher capacity. This is owing to the fact that the self-aggregation of the unique nanooctahedron-in-nanofiber structure has been greatly reduced because of the attachment of nanopolyhedrons in the long nanofibers. This unique microstructured cathode results in the large effective contact areas of the active materials, conductive additives and fully realize the advantage of nanomaterial-based cathodes.

Extraction of latent patterns and contexts from social honest signals using hierarchical Dirichlet processes

A fundamental task in pervasive computing is reliable acquisition of contexts from sensor data. This is crucial to the operation of smart pervasive systems and services so that they might behave efficiently and appropriately upon a given context. Simple forms of context can often be extracted directly from raw data. Equally important, or more, is the hidden context and pattern buried inside the data, which is more challenging to discover. Most of existing approaches borrow methods and techniques from machine learning, dominantly employ parametric unsupervised learning and clustering techniques. Being parametric, a severe drawback of these methods is the requirement to specify the number of latent patterns in advance. In this paper, we explore the use of Bayesian nonparametric methods, a recent data modelling framework in machine learning, to infer latent patterns from sensor data acquired in a pervasive setting. Under this formalism, nonparametric prior distributions are used for data generative process, and thus, they allow the number of latent patterns to be learned automatically and grow with the data - as more data comes in, the model complexity can grow to explain new and unseen patterns. In particular, we make use of the hierarchical Dirichlet processes (HDP) to infer atomic activities and interaction patterns from honest signals collected from sociometric badges. We show how data from these sensors can be represented and learned with HDP. We illustrate insights into atomic patterns learned by the model and use them to achieve high-performance clustering. We also demonstrate the framework on the popular Reality Mining dataset, illustrating the ability of the model to automatically infer typical social groups in this dataset. Finally, our framework is generic and applicable to a much wider range of problems in pervasive computing where one needs to infer high-level, latent patterns and contexts from sensor data.