Maximizing Heterogeneous Server Utilization with Limited Availability Times for Divisible Loads Scheduling on Networked Systems.

Most of the available divisible-load scheduling models assume that all servers in networked systems are idle before workloads arrive and that they can remain available online during workload computation. In fact, this assumption is not always valid. Different servers on networked systems may have heterogenous available times. If we ignore the availability constraints when dividing and distributing workloads among servers, some servers may not be able to start processing their assigned load fractions or deliver them on time. In view of this, we propose a new multi-installment scheduling model based on server availability time constraints. To solve this problem, we design an efficient heuristic algorithm consisting of a repair strategy and a local search strategy, by which an optimal load partitioning scheme is derived. The repair strategy guarantees time constraints, while the local search strategy achieves optimality. We evaluate the performance via rigorous simulation experiments and our results show that the proposed algorithm is suitable for solving large-scale scheduling problems employing heterogeneous servers with arbitrary available times. The proposed algorithm is shown to be superior to the existing algorithm in terms of achieving a shorter makespan of workloads.

Self-supervised Assisted Active Learning for Skin Lesion Segmentation.

Label scarcity has been a long-standing issue for biomedical image segmentation, due to high annotation costs and professional requirements. Recently, active learning (AL) strategies strive to reduce annotation costs by querying a small portion of data for annotation, receiving much traction in the field of medical imaging. However, most of the existing AL methods have to initialize models with some randomly selected samples followed by active selection based on various criteria, such as uncertainty and diversity. Such random-start initialization methods inevitably introduce under-value redundant samples and unnecessary annotation costs. For the purpose of addressing the issue, we propose a novel self-supervised assisted active learning framework in the cold-start setting, in which the segmentation model is first warmed up with self-supervised learning (SSL), and then SSL features are used for sample selection via latent feature clustering without accessing labels. We assess our proposed methodology on skin lesions segmentation task. Extensive experiments demonstrate that our approach is capable of achieving promising performance with substantial improvements over existing baselines. Clinical Relevance- The proposed method can smartly select samples to annotate without requiring labels for model initialization, which can save annotation costs in clinical practice.

CETSA Feature Based Clustering for Protein Outlier Discovery by Protein-to-Protein Interaction Prediction.

The Cellular Thermal Shift Assay (CETSA) is a biophysical assay based on the principle of ligand-induced thermal stabilization of target proteins. This technology has revolutionized cell-based target engagement studies and has been used as guidance for drug design. Although many ap-plications of CETSA data have been explored, the correlations between CETSA data and protein-protein interactions (PPI) have barely been touched. In this study, we conduct the first exploration study applying CETSA data for PPI prediction. We use a machine learning method, Decision Tree, to predict PPI scores using proteins' CETSA features. It shows promising results that the predicted PPI scores closely match the ground-truth PPI scores. Furthermore, for a small number of protein pairs, whose PPI score predictions mismatch the ground truth, we use iterative clustering strategy to gradually reduce the number of these pairs. At the end of iterative clustering, the remaining protein pairs may have some unusual properties and are of scientific value for further biological investigation. Our study has demonstrated that PPI is a brand-new application of CETSA data. At the same time, it also manifests that CETSA data can be used as a new data source for PPI exploration study.

Towards high performance computing for molecular structure prediction using IBM Cell Broadband Engine--an implementation perspective.

BACKGROUND: RNA structure prediction problem is a computationally complex task, especially with pseudo-knots. The problem is well-studied in existing literature and predominantly uses highly coupled Dynamic Programming (DP) solutions. The problem scale and complexity become embarrassingly humungous to handle as sequence size increases. This makes the case for parallelization. Parallelization can be achieved by way of networked platforms (clusters, grids, etc) as well as using modern day multi-core chips. METHODS: In this paper, we exploit the parallelism capabilities of the IBM Cell Broadband Engine to parallelize an existing Dynamic Programming (DP) algorithm for RNA secondary structure prediction. We design three different implementation strategies that exploit the inherent data, code and/or hybrid parallelism, referred to as C-Par, D-Par and H-Par, and analyze their performances. Our approach attempts to introduce parallelism in critical sections of the algorithm. We ran our experiments on SONY Play Station 3 (PS3), which is based on the IBM Cell chip. RESULTS: Our results suggest that introducing parallelism in DP algorithm allows it to easily handle longer sequences which otherwise would consume a large amount of time in single core computers. The results further demonstrate the speed-up gain achieved in exploiting the inherent parallelism in the problem and also elicits the advantages of using multi-core platforms towards designing more sophisticated methodologies for handling a fairly long sequence of RNA. CONCLUSION: The speed-up performance reported here is promising, especially when sequence length is long. To the best of our literature survey, the work reported in this paper is probably the first-of-its-kind to utilize the IBM Cell Broadband Engine (a heterogeneous multi-core chip) to implement a DP. The results also encourage using multi-core platforms towards designing more sophisticated methodologies for handling a fairly long sequence of RNA to predict its secondary structure.

Deeply Supervised Active Learning for Finger Bones Segmentation.

Segmentation is a prerequisite yet challenging task for medical image analysis. In this paper, we introduce a novel deeply supervised active learning approach for finger bones segmentation. The proposed architecture is fine-tuned in an iterative and incremental learning manner. In each step, the deep supervision mechanism guides the learning process of hidden layers and selects samples to be labeled. Extensive experiments demonstrated that our method achieves competitive segmentation results using less labeled samples as compared with full annotation.Clinical relevance- The proposed method only needs a few annotated samples on the finger bones task to achieve comparable results in comparison with full annotation, which can be used to segment finger bones for medical practices, and generalized into other clinical applications.

Multi-Phase Cross-modal Learning for Noninvasive Gene Mutation Prediction in Hepatocellular Carcinoma.

Hepatocellular carcinoma (HCC) is the most common type of primary liver cancer and the fourth most common cause of cancer-related death worldwide. Understanding the underlying gene mutations in HCC provides great prognostic value for treatment planning and targeted therapy. Radiogenomics has revealed an association between non-invasive imaging features and molecular genomics. However, imaging feature identification is laborious and error-prone. In this paper, we propose an end-to-end deep learning framework for mutation prediction in APOB, COL11A1 and ATRX genes using multiphasic CT scans. Considering intra-tumour heterogeneity (ITH) in HCC, multi-region sampling technology is implemented to generate the dataset for experiments. Experimental results demonstrate the effectiveness of the proposed model.

Predicting Stroke from Electronic Health Records.

Studies have identified various risk factors associated with the onset of stroke in an individual. Data mining techniques have been used to predict the occurrence of stroke based on these factors by using patients' medical records. However, there has been limited use of electronic health records to study the inter-dependency of different risk factors of stroke. In this paper, we perform an analysis of patients' electronic health records to identify the impact of risk factors on stroke prediction. We also provide benchmark performance of the state-of-art machine learning algorithms for predicting stroke using electronic health records.

CardioXNet: Automated Detection for Cardiomegaly Based on Deep Learning.

In this paper, we present an automated procedure to determine the presence of cardiomegaly on chest X-ray image based on deep learning. The proposed algorithm CardioXNet uses deep learning methods U-NET and cardiothoracic ratio for diagnosis of cardiomegaly from chest X-rays. U-NET learns the segmentation task from the ground truth data. OpenCV is used to denoise and maintain the precision of region of interest once minor errors occur. Therefore, Cardiothoracic ratio (CTR) is calculated as a criterion to determine cardiomegaly from U-net segmentations. End-to-end Dense-Net neural network is used as baseline. This study has shown that the feasibility of combing deep learning segmentation and medical criterion to automatically recognize heart disease in medical images with high accuracy and agreement with the clinical results.

Automated Classification Using End-to-End Deep Learning.

According to a study [1] by the Ministry of Health in Singapore, since 2009, cancer, ischaemic heart disease and pneumonia together accounted for approximately 60% of the total causes of death. It has been 9 years, and Pneumonia and other Acute Upper Respiratory Infections still is one of the top 10 conditions of hospitalization. In cases of respiratory diseases such as Chronic Obstructive Pulmonary Disease (COPD), it has been found that close to 55% of cases are misdiagnosed. This is shocking as, an early diagnosis of respiratory diseases can lead to an earlier treatment intervention, ultimately lessening symptoms, slowing the progression, and improving overall quality of life. With the advent of Deep Neural Network architectures which have shown phenomenal results in the field of Computer Assisted Diagnosis (CAD), we hope to implement a Lung Classification Model using End-to-End Deep learning to classify Chest X-Ray images into one of 14 primary classes of lung diseases. Using our implementation of the Densely Connected Convolutional Neural Network model architecture, we aim to increase existing model accuracy in Lung Disease classification by iteratively reducing the search space and region of interest for different. We shall experiment on a 14-class classification model and compare the results with a binary classifier as well, to understand the performance of DenseNets on Chest X-Ray (CXR) Data with a reduced search space.

Aligning biological sequences on distributed bus networks: a divisible load scheduling approach.

In this paper, we design a multiprocessor strategy that exploits the computational characteristics of the algorithms used for biological sequence comparison proposed in the literature. We employ divisible load theory (DLT) that is suitable for handling large scale processing on network based systems. For the first time in the domain of DLT, the problem of aligning biological sequences is attempted. The objective is to minimize the total processing time of the alignment process. In designing our strategy, DLT facilitates a clever partitioning of the entire computation process involved in such a way that the overall time consumed for aligning the sequences is a minimum. The partitioning takes into account the computation speeds of the nodes and the underlying communication network. Since this is a real-life application, the post-processing phase becomes important, and hence we consider propagating the results back in order to generate an exact alignment. We consider several cases in our analysis such as deriving closed-form solutions for the processing time for heterogeneous, homogeneous, and networks with slow links. Further, we attempt to employ a multiinstallment strategy to distribute the tasks such that a higher degree of parallelism can be achieved. For slow networks, our strategy recommends near-optimal solutions. We derive an important condition to identify such cases and propose two heuristic strategies. Also, our strategy can be extended for multisequence alignment by utilizing a clustering strategy such as the Berger-Munson algorithm proposed in the literature. Finally, we use real-life DNA samples of house mouse mitochondrion (Mus Musculus Mitochondrion, NC_001569) consisting of 16,295 residues and the DNA of human mitochondrion (Homo Sapiens Mitochondrion, NC_001807) consisting of 16,571 residues, obtainable from the GenBank, in our rigorous simulation experiments to illustrate all the theoretical findings.

Inferring transcription factor interactions using a novel HV-SVM classifier.

Interactions between Transcription Factors (TFs) are necessary for deciphering the complex mechanisms of transcription regulation in eukaryotes. We proposed a novel HV-kernel based SVM classifier to classify TF-TF pairs based on their protein domains and GO annotations. Two types of pairwise kernels, namely, a horizontal kernel and a vertical kernel, were combined to evaluate the similarity between a pair of TFs, and a Genetic Algorithm was used to obtain kernel and feature weights to optimise the classifier's performance. We showed that our proposed HV-SVM method can make accurate predictions of TF-TF interactions even in the higher and more complex eukaryotes.