Nadine Hajj

Weighted entropy cortical algorithms for isolated Arabic speech recognition

Cortical algorithms (CA) inspired by and modeled after the human cortex, have shown superior accuracy in few machine learning applications. However, CA have not been extensively implemented for speech recognition applications, in particular the Arabic language. Motivated to apply CA to Arabic speech recognition, we present in this paper an improved CA that is efficiently trained using an entropy-based cost function, and an entropy based weight update rule. We modify the strengthening and inhibiting rules originally employed in CA during feedback training with weighted entropy concepts. Preliminary results show the merit of the proposed modifications in the recognition of isolated Arabic speech and motivate follow on research.

A MapReduce Cortical Algorithms Implementation for Unsupervised Learning of Big Data

In the big data era, the need for fast robust machine learning techniques is rapidly increasing. Deep network architectures such as cortical algorithms are challenged by big data problems which result in lengthy and complex training. In this paper, we present a distributed cortical algorithm implementation for the unsupervised learning of big data based on a combined node-data parallelization scheme. A data sparsity measure is used to divide the data before distributing the columns in the network over many computing nodes based on the MapReduce framework. Experimental results on multiple datasets showed an average speedup of 8.1× compared to serial implementations.

Use of Contrast-Enhanced Ultrasound for Confirmation of Central Venous Catheter Placement: Systematic Review and Meta-analysis

The purpose of this meta‐analysis was to determine the sensitivity, specificity, and positive and negative predictive values of contrast‐enhanced ultrasound (US) for confirming the tip location and placement of central venous catheters in adult patients.

Isolated handwriting recognition via multi-stage Support Vector Machines

Since isolated letter handwriting recognition is an essential step for online hand writing recognition, we present in this paper an efficient and writer independent isolated letter handwriting recognition system using pen trajectory modeling for feature extraction and a multi-stage Support Vector Machines (SVM) for classification. Inheriting the good discriminating ability of SVM while modeling sequential data, this hierarchical approach shows using 4 fold validation an average accuracy of 91.8% on the UJIpenchars database that consists of a collection of 1144 isolated letters written by 11 different writers. To the best of our knowledge, the best recognition rate achieved on this database is 89.15% using Dynamic Time Wrapping and 3 nearest neighbor classifier.

A subjectivity classification framework for sports articles using improved cortical algorithms

AbstractThe enormous number of articles published daily on the Internet, by a diverse array of authors, often offers misleading or unwanted information, rendering activities such as sports betting riskier. As a result, extracting meaningful and reliable information from these sources becomes a time-consuming and near impossible task. In this context, labeling articles as objective or subjective is not a simple natural language processing task because subjectivity can take several forms. With the rise of online sports betting due to the revolution in Internet and mobile technology, an automated system capable of sifting through all these data and finding relevant sources in a reasonable amount of time presents itself as a desirable and marketable product. In this work, we present a framework for the classification of sports articles composed of three stages: The first stage extracts articles from web pages using text extraction libraries, parses the text and then tags words using Stanford’s parts of speech tagger; the second stage extracts unique syntactic and semantic features, and reduces them using our modified cortical algorithm (CA)—hereafter CA*—while the third stage classifies these texts as objective or subjective. Our framework was tested on a database containing 1000 articles, manually labeled using Amazon’s crowdsourcing tool, Mechanical Turk; and results using CA, CA*, support vector machines and one of its soft computing variants (LMSVM) as classifiers were reported. A testing accuracy of 85.6% was achieved on a fourfold cross-validation with a 40% reduction in features using CA* that was trained using an entropy weight update rule and a cross-entropy cost function.

Outcomes of patients with systolic heart failure presenting with sepsis to the emergency department of a tertiary hospital: a retrospective chart review study from Lebanon

Objectives Patients with congestive heart failure (CHF) may be at a higher risk of mortality from sepsis than patients without CHF due to insufficient cardiovascular reserves during systemic infections. The aim of this study is to compare sepsis-related mortality between CHF and no CHF in patients presenting to a tertiary medical centre. Design A single-centre, retrospective, cohort study. Setting Conducted in an academic emergency department (ED) between January 2010 and January 2015. Patients’ charts were queried via the hospital’s electronic system. Patients with a diagnosis of sepsis were included. Descriptive analysis was performed on the demographics, characteristics and outcomes of patients with sepsis of the study population. Participants A total of 174 patients, of which 87 (50%) were patients with CHF. Primary and secondary outcomes The primary outcome of the study was in-hospital mortality. Secondary outcomes included intensive care unit (ICU) and hospital lengths of stay, and differences in interventions between the two groups. Results Patients with CHF had a higher in-hospital mortality (57.5% vs 34.5%). Patients with sepsis and CHF had higher odds of death compared with the control population (OR 2.45; 95% CI 1.22 to 4.88). Secondary analyses showed that patients with CHF had lower instances of bacteraemia on presentation to the ED (31.8% vs 46.4%). They had less intravenous fluid requirements in first 24 hours (2.75±2.28 L vs 3.67±2.82 L, p =0.038), had a higher rate of intubation in the ED (24.2% vs 10.6%, p=0.025) and required more dobutamine in the first 24 hours (16.1% vs 1.1%, p<0.001). ED length of stay was found to be lower in patients with CHF (15.12±24.45 hours vs 18.17±26.13 hours, p=0.418) and they were more likely to be admitted to the ICU (59.8% vs 48.8%, p=0.149). Conclusion Patients with sepsis and CHF experienced an increased hospital mortality compared with patients without CHF.

A piecewise weight update rule for a supervised training of cortical algorithms

First introduced by MountCastle, cortical algorithms (CA) are positioned to outperform artificial neural networks second generations due to their ability to hierarchically store sequences of patterns in an invariant form. Despite their closer resemblance to the human cortex and their hypothetical improved performance, CA adoption as a deep learning approach remains limited in energy aware environments due to their high computational training complexity. Motivated to reduce CA supervised training complexity in limited hardware resources environments, we propose in this paper a piecewise linear or polygonal weight update rule for a supervised training of CA based on a linearization of the exponential function. As shown by our simulation results on 12 publicly available databases and our developed error-bound proofs, the proposed rule reduces CA training time by a factor of 3 at the expense of a 0.5% degradation in accuracy. A simpler approximation relying on the asymptotes at 0 and infinity reduces training time by a factor of 3.5 coupled with a reduction of 1.49% in accuracy.

A Game Theory Approach to Demand Side Management in Smart Grids

With the increase in global energy awareness, smart grids improve the efficiency and peak leveling of power systems. Demand side management is the controlling scheme in such grids and it aims to optimize several characteristics using an interactive dynamic pricing scheme. In this paper we propose a game theoretic approach to the demand side management where several subscribers share one common energy supplier. In our model, users send their demand vectors for an upcoming period of time to the network and the energy provider responds by broadcasting a dynamic price vector. Energy consumers are concerned with minimizing their total energy cost per day while the power provider aims to maximize its profit while minimizing the peak to average load ratio. Converging to a unique Nash equilibrium solution using a dual constrained optimization problem, our model results motivate follow on research.