Jianjia Zhang

Unintended adverse consequences of introducing electronic health records in residential aged care homes

PURPOSE The aim of this study was to investigate the unintended adverse consequences of introducing electronic health records (EHR) in residential aged care homes (RACHs) and to examine the causes of these unintended adverse consequences. METHOD A qualitative interview study was conducted in nine RACHs belonging to three organisations in the Australian Capital Territory (ACT), New South Wales (NSW) and Queensland, Australia. A longitudinal investigation after the implementation of the aged care EHR systems was conducted at two data points: January 2009 to December 2009 and December 2010 to February 2011. Semi-structured interviews were conducted with 110 care staff members identified through convenience sampling, representing all levels of care staff who worked in these facilities. Data analysis was guided by DeLone and McLean Information Systems Success Model, in reference with the previous studies of unintended consequences for the introduction of computerised provider order entry systems in hospitals. RESULTS Eight categories of unintended adverse consequences emerged from 266 data items mentioned by the interviewees. In descending order of the number and percentage of staff mentioning them, they are: inability/difficulty in data entry and information retrieval, end user resistance to using the system, increased complexity of information management, end user concerns about access, increased documentation burden, the reduction of communication, lack of space to place enough computers in the work place and increasing difficulties in delivering care services. The unintended consequences were caused by the initial conditions, the nature of the EHR system and the way the system was implemented and used by nursing staff members. CONCLUSIONS Although the benefits of the EHR systems were obvious, as found by our previous study, introducing EHR systems in RACH can also cause adverse consequences of EHR avoidance, difficulty in access, increased complexity in information management, increased documentation burden, reduction of communication and the risks of lacking care follow-up, which may cause negative effects on aged care services. Further research can focus on investigating how the unintended adverse consequences can be mitigated or eliminated by understanding more about nursing staffs work as well as the information flow in RACH. This will help to improve the design, introduction and management of EHR systems in this setting.

HEp-2 Cell Image Classification With Deep Convolutional Neural Networks

Efficient Human Epithelial-2 cell image classification can facilitate the diagnosis of many autoimmune diseases. This paper proposes an automatic framework for this classification task, by utilizing the deep convolutional neural networks (CNNs) which have recently attracted intensive attention in visual recognition. In addition to describing the proposed classification framework, this paper elaborates several interesting observations and findings obtained by our investigation. They include the important factors that impact network design and training, the role of rotation-based data augmentation for cell images, the effectiveness of cell image masks for classification, and the adaptability of the CNN-based classification system across different datasets. Extensive experimental study is conducted to verify the above findings and compares the proposed framework with the well-established image classification models in the literature. The results on benchmark datasets demonstrate that 1) the proposed framework can effectively outperform existing models by properly applying data augmentation, 2) our CNN-based framework has excellent adaptability across different datasets, which is highly desirable for cell image classification under varying laboratory settings. Our system is ranked high in the cell image classification competition hosted by ICPR 2014.

Beyond Covariance: Feature Representation with Nonlinear Kernel Matrices

Covariance matrix has recently received increasing attention in computer vision by leveraging Riemannian geometry of symmetric positive-definite (SPD) matrices. Originally proposed as a region descriptor, it has now been used as a generic representation in various recognition tasks. However, covariance matrix has shortcomings such as being prone to be singular, limited capability in modeling complicated feature relationship, and having a fixed form of representation. This paper argues that more appropriate SPD-matrix-based representations shall be explored to achieve better recognition. It proposes an open framework to use the kernel matrix over feature dimensions as a generic representation and discusses its properties and advantages. The proposed framework significantly elevates covariance representation to the unlimited opportunities provided by this new representation. Experimental study shows that this representation consistently outperforms its covariance counterpart on various visual recognition tasks. In particular, it achieves significant improvement on skeleton-based human action recognition, demonstrating the state-of-the-art performance over both the covariance and the existing non-covariance representations.

Functional Brain Network Classification With Compact Representation of SICE Matrices

Recently, a sparse inverse covariance estimation (SICE) technique has been employed to model functional brain connectivity. The inverse covariance matrix (SICE matrix in short) estimated for each subject is used as a representation of brain connectivity to discriminate Alzheimers disease from normal controls. However, we observed that direct use of the SICE matrix does not necessarily give satisfying discrimination, due to its high dimensionality and the scarcity of training subjects. Looking into this problem, we argue that the intrinsic dimensionality of these SICE matrices shall be much lower, considering 1) an SICE matrix resides on a Riemannian manifold of symmetric positive definiteness matrices, and 2) human brains share common patterns of connectivity across subjects. Therefore, we propose to employ manifold-based similarity measures and kernel-based PCA to extract principal connectivity components as a compact representation of brain network. Moreover, to cater for the requirement of both discrimination and interpretation in neuroimage analysis, we develop a novel preimage estimation algorithm to make the obtained connectivity components anatomically interpretable. To verify the efficacy of our method and gain insights into SICE-based brain networks, we conduct extensive experimental study on synthetic data and real rs-fMRI data from the ADNI dataset. Our method outperforms the comparable methods and improves the classification accuracy significantly.

Learning Discriminative Stein Kernel for SPD Matrices and Its Applications

Stein kernel (SK) has recently shown promising performance on classifying images represented by symmetric positive definite (SPD) matrices. It evaluates the similarity between two SPD matrices through their eigenvalues. In this paper, we argue that directly using the original eigenvalues may be problematic because: 1) eigenvalue estimation becomes biased when the number of samples is inadequate, which may lead to unreliable kernel evaluation, and 2) more importantly, eigenvalues reflect only the property of an individual SPD matrix. They are not necessarily optimal for computing SK when the goal is to discriminate different classes of SPD matrices. To address the two issues, we propose a discriminative SK (DSK), in which an extra parameter vector is defined to adjust the eigenvalues of input SPD matrices. The optimal parameter values are sought by optimizing a proxy of classification performance. To show the generality of the proposed method, three kernel learning criteria that are commonly used in the literature are employed as a proxy. A comprehensive experimental study is conducted on a variety of image classification tasks to compare the proposed DSK with the original SK and other methods for evaluating the similarity between SPD matrices. The results demonstrate that the DSK can attain greater discrimination and better align with classification tasks by altering the eigenvalues. This makes it produce higher classification performance than the original SK and other commonly used methods.

Revisiting Metric Learning for SPD Matrix Based Visual Representation

The success of many visual recognition tasks largely depends on a good similarity measure, and distance metric learning plays an important role in this regard. Meanwhile, Symmetric Positive Definite (SPD) matrix is receiving increased attention for feature representation in multiple computer vision applications. However, distance metric learning on SPD matrices has not been sufficiently researched. A few existing works approached this by learning either d2 × p or d × k transformation matrix for d× d SPD matrices. Different from these methods, this paper proposes a new member to the family of distance metric learning for SPD matrices. It learns only d parameters to adjust the eigenvalues of the SPD matrices through an efficient optimisation scheme. Also, it is shown that the proposed method can be interpreted as learning a sample-specific transformation matrix, instead of the fixed transformation matrix learned for all the samples in the existing works. The optimised d parameters can be used to massage the SPD matrices for better discrimination while still keeping them in the original space. From this perspective, the proposed method complements, rather than competes with, the existing linear-transformation-based methods, as the latter can always be applied to the output of the former to perform distance metric learning in further. The proposed method has been tested on multiple SPD-based visual representation data sets used in the literature, and the results demonstrate its interesting properties and attractive performance.

Exploring Compact Representation of SICE Matrices for Functional Brain Network Classification

Recently, sparse inverse covariance matrix (SICE matrix) has been used as a representation of brain connectivity to classify Alzheimer’s disease and normal controls. However, its high dimensionality can adversely affect the classification performance. Considering the underlying manifold where SICE matrices reside and the common patterns shared by brain connectivity across subjects, we propose to explore the lower dimensional intrinsic components of SICE matrix for compact representation. This leads to significant improvements of brain connectivity classification. Moreover, to cater for the requirement of both discrimination and interpretation in neuroimage analysis, we develop a novel pre-image estimation algorithm to make the obtained connectivity components anatomically interpretable. The advantages of our method have been well demonstrated on both synthetic and real rs-fMRI data sets.

Subject-adaptive Integration of Multiple SICE Brain Networks with Different Sparsity

Abstract As a principled method for partial correlation estimation, sparse inverse covariance estimation (SICE) has been employed to model brain connectivity networks, which holds great promise for brain disease diagnosis. For each subject, the SICE method naturally leads to a set of connectivity networks with various sparsity. However, existing methods usually select a single network from them for classification and the discriminative power of this set of networks has not been fully exploited. This paper argues that the connectivity networks at different sparsity levels present complementary connectivity patterns and therefore they should be jointly considered to achieve high classification performance. In this paper, we propose a subject-adaptive method to integrate multiple SICE networks as a unified representation for classification. The integration weight is learned adaptively for each subject in order to endow the method with the flexibility in dealing with subject variations. Furthermore, to respect the manifold geometry of SICE networks, Stein kernel is employed to embed the manifold structure into a kernel-induced feature space, which allows a linear integration of SICE networks to be designed. The optimization of the integration weight and the classification of the integrated networks are performed via a sparse representation framework. Through our method, we provide a unified and effective network representation that is transparent to the sparsity level of SICE networks, and can be readily utilized for further medical analysis. Experimental study on ADHD and ADNI data sets demonstrates that the proposed integration method achieves notable improvement of classification performance in comparison with methods using a single sparsity level of SICE networks and other commonly used integration methods, such as Multiple Kernel Learning.

Accelerating the Divisive Information-Theoretic Clustering of Visual Words

Word clustering is an effective approach in the bag- of-words model to reducing the dimensionality of high-dimensional features. In recent years, the bag- of-words model has been successfully introduced into visual recognition and significantly developed. Often, in order to adequately model the complex and diversified visual patterns, a large number of visual words are used, especially in the state-of- the-art visual recognition methods. As a result, the existing word clustering algorithms become not computationally efficient enough. They can considerably prolong the process such as model updating and parameter tuning, where word clustering needs to be repeatedly employed. In this paper, we focus on the divisive information-theoretic clustering, one of the most efficient word clustering algorithms in the field of text analysis, and accelerate its speed to better deal with a large number of visual words. We discuss the properties of its cluster membership evaluation function, KL- divergence, in both binary and multi-class classification cases and develop the accelerated versions in two different ways. Theoretical analysis shows that the proposed accelerated divisive information-theoretic clustering algorithm can handle a large number of visual words in a much more efficient manner. As demonstrated on the benchmark datasets in visual recognition, it can achieve speed-up by hundreds of times while well maintaining the clustering performance of the original algorithm.

A Fast Approximate AIB Algorithm for Distributional Word Clustering

Distributional word clustering merges the words having similar probability distributions to attain reliable parameter estimation, compact classification models and even better classification performance. Agglomerative Information Bottleneck (AIB) is one of the typical word clustering algorithms and has been applied to both traditional text classification and recent image recognition. Although enjoying theoretical elegance, AIB has one main issue on its computational efficiency, especially when clustering a large number of words. Different from existing solutions to this issue, we analyze the characteristics of its objective function-the loss of mutual information, and show that by merely using the ratio of word-class joint probabilities of each word, good candidate word pairs for merging can be easily identified. Based on this finding, we propose a fast approximate AIB algorithm and show that it can significantly improve the computational efficiency of AIB while well maintaining or even slightly increasing its classification performance. Experimental study on both text and image classification benchmark data sets shows that our algorithm can achieve more than 100 times speedup on large real data sets over the state-of-the-art method.