Tree-Structured Stick Breaking Processes for Hierarchical Data

Many data are naturally modeled by an unobserved hierarchical structure. In this paper we propose a flexible nonparametric prior over unknown data hierarchies. The approach uses nested stick-breaking processes to allow for trees of unbounded width and depth, where data can live at any node and are infinitely exchangeable. One can view our model as providing infinite mixtures where the components have a dependency structure corresponding to an evolutionary diffusion down a tree. By using a stick-breaking approach, we can apply Markov chain Monte Carlo methods based on slice sampling to perform Bayesian inference and simulate from the posterior distribution on trees. We apply our method to hierarchical clustering of images and topic modeling of text data.

Development of a universal stress sensor for graphene and carbon fibres

Carbon fibres are a significant volume fraction of modern structural airframes. Embedded into polymer matrices, they provide significant strength and stiffness gains by unit weight compared with competing structural materials. Here we use the Raman G peak to assess the response of carbon fibres to the application of strain, with reference to the response of graphene itself. Our data highlight the predominance of the in-plane graphene properties in all graphitic structures examined. A universal master plot relating the G peak strain sensitivity to tensile modulus of all types of carbon fibres, as well as graphene, is presented. We derive a universal value of - average - phonon shift rate with axial stress of around -5ω0 -1 (cm -1 Mpa-1), where ω0 is the G peak position at zero stress for both graphene and carbon fibre with annular morphology. The use of this for stress measurements in a variety of applications is discussed. © 2011 Macmillan Publishers Limited. All rights reserved.

Modelling cognitive impairment to improve universal access

The purpose of this study is to develop a model of cognitive impairment to help designers consider the range of issues which affect the lives of people living with such impairment. A series of interviews with experts of cognitive impairment was conducted to describe and assess the links between specific medical conditions, including learning disability, specific learning difficulties, autistic spectrum disorders, traumatic brain injury and schizophrenia, and the types of cognitive impairment associated with them. The results reveal some of the most prevalent and serious types of impairment, which - when transformed into design guidance - will help designers make mainstream products more inclusive also for people with cognitive impairment. © 2011 Springer-Verlag.

Decision tree-based context clustering based on cross validation and hierarchical priors

The standard, ad-hoc stopping criteria used in decision tree-based context clustering are known to be sub-optimal and require parameters to be tuned. This paper proposes a new approach for decision tree-based context clustering based on cross validation and hierarchical priors. Combination of cross validation and hierarchical priors within decision tree-based context clustering offers better model selection and more robust parameter estimation than conventional approaches, with no tuning parameters. Experimental results on HMM-based speech synthesis show that the proposed approach achieved significant improvements in naturalness of synthesized speech over the conventional approaches. © 2011 IEEE.

Semi-supervised video segmentation using tree structured graphical models

We present a novel, implementation friendly and occlusion aware semi-supervised video segmentation algorithm using tree structured graphical models, which delivers pixel labels alongwith their uncertainty estimates. Our motivation to employ supervision is to tackle a task-specific segmentation problem where the semantic objects are pre-defined by the user. The video model we propose for this problem is based on a tree structured approximation of a patch based undirected mixture model, which includes a novel time-series and a soft label Random Forest classifier participating in a feedback mechanism. We demonstrate the efficacy of our model in cutting out foreground objects and multi-class segmentation problems in lengthy and complex road scene sequences. Our results have wide applicability, including harvesting labelled video data for training discriminative models, shape/pose/articulation learning and large scale statistical analysis to develop priors for video segmentation. © 2011 IEEE.