The mobile health (mHealth) system, leveraging mobile edge computing, can monitor health status and provide diagnosis. However, due to the privacy of medical data and the resource limitations of mobile devices, patients are unable to access diagnostic services provided by untrusted servers in real-time. Existing schemes present significant challenges in private heterogeneous data aggregation, model training and inference in the presence of malicious participants, and expensive resource consumption. To address these issues, in this paper, we propose a non-interactive privacy-preserving system with the naive Bayesian model, i.e.,$\mathsf {RobustHealth}$, for heterogeneous mHealth diagnosis. Specifically, we extract homogeneous features from heterogeneous datasets to enable efficient encrypted aggregation. We propose a novel private model training algorithm with enhanced security to against collusion-then-differential attacks. We develop a novel non-interactive private model inference algorithm using minimal lightweight cryptographic primitives, designed for patients under unstable network environments. We provide formal security proofs for our system using the Universal Composable (UC) framework. To validate the performance of$\mathsf {RobustHealth}$, we conduct extensive experiments on real-world heterogeneous datasets, and compared with related works. The results demonstrate a$\bf {4.37\%}$improvement in model accuracy, along with significant reductions in computational and communication overheads of$\bf {21.18\times }$and$\bf {4.24\times }$, respectively.

Most products at ByteDance, e.g., TikTok, Douyin, and Toutiao, naturally generate massive amounts of graph data. To efficiently store, query and update massive graph data is challenging for the broad range of products at ByteDance with various performance requirements. We categorize graph workloads at ByteDance into three types: online analytical, transaction, and serving processing, where each workload has its own characteristics. Existing graph databases have different performance bottlenecks in handling these workloads and none can efficiently handle the scale of graphs at ByteDance. We developed ByteGraph to process these graph workloads with high throughput, low latency and high scalability. There are several key designs in ByteGraph that make it efficient for processing our workloads, including edge-trees to store adjacency lists for high parallelism and low memory usage, adaptive optimizations on thread pools and indexes, and geographic replications to achieve fault tolerance and availability. ByteGraph has been in production use for several years and its performance has shown to be robust for processing a wide range of graph workloads at ByteDance.

Multilevel cell (MLC) NAND flash memory uses the voltages of the memory cells to represent bits, but high voltages cause much more damage on the cells than low voltages. Free space in MLC can be leveraged to reduce the usage of the high voltages and thus extend the lifetime of MLC. However, limited by the conventional data representation rule that represents bits by the voltage of one single cell, the high voltages are used in a high probability. To fully explore the potential of the free space on reducing the usage of high voltages without changing the total density of MLC, we propose a novel data representation aware of damage, named DREAM. DREAM uses the low voltage combinations of multiple cells instead of the voltage of one single cell to represent bits. It enables to represent the same bits through flexibly replacing the high voltages in some cells with the low voltages in other cells when free space is available. Hence, high voltages which cause more damage are less used and the lifetime of the MLC memory is extended. In addition, complementary techniques are proposed to mitigate performance loss induced by DREAM. Theoretical analysis and simulation results demonstrate the effectiveness and efficiency of DREAM.

Effective and safe implementation of precision oncology for breast cancer is a vital strategy to improve patient outcomes, which relies on the application of reliable biomarkers. As 'liquid biopsy' and novel resource for biomarkers, exosomes provide a promising avenue for the diagnosis and treatment of breast cancer. Although several exosome-related databases have been developed, there is still lacking of an integrated database for exosome-based biomarker discovery. To this end, a comprehensive database ExoBCD (https://exobcd.liumwei.org) was constructed with the combination of robust analysis of four high-throughput datasets, transcriptome validation of 1191 TCGA cases and manual mining of 950 studies. In ExoBCD, approximately 20 900 annotation entries were integrated from 25 external sources and 306 exosomal molecules (49 potential biomarkers and 257 biologically interesting molecules). The latter could be divided into 3 molecule types, including 121 mRNAs, 172 miRNAs and 13 lncRNAs. Thus, the well-linked information about molecular characters, experimental biology, gene expression patterns, overall survival, functional evidence, tumour stage and clinical use were fully integrated. As a data-driven and literature-based paradigm proposed of biomarker discovery, this study also demonstrated the corroborative analysis and identified 36 promising molecules, as well as the most promising prognostic biomarkers, IGF1R and FRS2. Taken together, ExoBCD is the first well-corroborated knowledge base for exosomal studies of breast cancer. It not only lays a foundation for subsequent studies but also strengthens the studies of probing molecular mechanisms, discovering biomarkers and developing meaningful clinical use.

In tumor metastasis, the margination and adhesion of tumor cells are two critical and closely related steps, which may determine the destination where the tumor cells extravasate to. We performed a direct three-dimensional simulation on the behaviors of the tumor cells in a real microvascular network, by a hybrid method of the smoothed dissipative particle dynamics and immersed boundary method (SDPD-IBM). The tumor cells are found to adhere at the microvascular bifurcations more frequently, and there is a positive correlation between the adhesion of the tumor cells and the wall-directed force from the surrounding red blood cells (RBCs). The larger the wall-directed force is, the closer the tumor cells are marginated towards the wall, and the higher the probability of adhesion behavior happen is. A relatively low or high hematocrit can help to prevent the adhesion of tumor cells, and similarly, increasing the shear rate of blood flow can serve the same purpose. These results suggest that the tumor cells may be more likely to extravasate at the microvascular bifurcations if the blood flow is slow and the hematocrit is moderate.

With the rapid development of biomedical technology, discovering causality from genes and human physiological and pathological characteristics has become a hot but challenge spot over the past decades. Due to the increment of the amount of biomedical data, discovering causality from observed data becomes more and more difficult to search this large body of knowledge in a meaningful manner. To address the issues in existing causality discovering models, we introduce a generative Bayesian causal network that combines neural network to explicitly characterize these unique causal-effect relationships as a variable number of nodes and links. Particularly, a basic skeleton is generated for node selection to reduce the network size by minimizing the maximum mean discrepancy among variables. In addition, a causal generative neural network model is presented to construct causal network with cause-effect scores between variables. Empirical evaluations on two publicly available biomedical datasets and four synthetic datasets suggest our approach significantly outperforms the state-of-the-art methods in discovering causal relationships among biomedical variables.

Power-aware task scheduling on processors has been a research hotspot in computing systems. Given an application G containing a set N of tasks {n1,...,n|N|}, and a system containing a set U of processors {u1,..., u|U|}, the power-aware task scheduling generally refers to finding the appropriate processor and frequency for each task ni, so as to make sure that all the tasks can be finished efficiently and the overall energy consumption is guaranteed. In this article, we study the problem of minimizing the schedule length for energy consumption constrained parallel applications on heterogeneous computing systems, where the schedule length refers to the time interval between starting the first task and finishing the last task. For this problem, existing work adopts a policy that preassigns the minimum energy consumption for each unassigned task. Nevertheless, our analysis reveals that, such a pre-assignment policy could be unfair for the low priority tasks, and it may not achieve an optimistic schedule length. Thereby, we propose a new task scheduling algorithm that suggests a weight-based mechanism to preassign energy consumption for unassigned tasks, and we provide the rigorous proof to show its feasibility. Further, we show that this idea can be extended to solve reliability maximization problems with energy consumption constraint or with both deadline and energy consumption constraints, where the reliability refers to the probability of executing application G without failures, and the deadline constraint refers to the “allowable” maximum schedule length. We have conducted extensive experiments based on real parallel applications. The experimental results consistently demonstrate that our proposed algorithms can achieve favourable performance, compared to state-of-the-art algorithms.

Power-aware task scheduling on processors has been a hot topic. In this paper, we study the problem of minimizing the schedule length for energy consumption constrained parallel applications on heterogeneous distributed systems. Previous work (solving this problem) adopts a policy that preassigns the minimum energy consumption for each unassigned task. Nevertheless, our analysis reveals that such a preassignment policy could be unfair, and it may not achieve an optimistic schedule length. Motivated by this, we propose a new task scheduling algorithm that suggests a weight-based mechanism to preassign energy consumption for unassigned tasks. We theoretically prove that our preassignment mechanism can guarantee the energy consumption constraint. Also, we have conducted extensive experiments based on two real parallel applications. The results consistently demonstrate that, compared to state-of-the-art algorithms, our approach can achieve smaller schedule length while satisfying the energy consumption constraint.

Relevance ranking models based on additive ensembles of regression trees have shown quite good effectiveness in web search engines. In the era of big data, tree ensemble models grow large in both tree depth and ensemble size to provide even better search relevance and user experience. However, the computational cost for their scoring process is high, such that it becomes a challenging issue to apply the big tree ensemble models in a search engine which needs to answer thousands of queries per second. Although several works have been proposed to improve the scoring process, the challenge is still great especially when the model size grows large. In this paper, we present RapidScorer , a novel framework for speeding up the scoring process of industry-scale tree ensemble models, without hurting the quality of scoring results. RapidScorer introduces a modified run length encoding called epitome to the bitvector representation of the tree nodes. Epitome can greatly reduce the computation cost to traverse the tree ensemble, and work with several other proposed strategies to maximize the compactness of data units in memory. The achieved compactness makes it possible to fully utilize data parallelization to improve model scalability. Experiments on two web search benchmarks show that, RapidScorer achieves significant speed-up over the state-of-the-art methods: V-QuickScorer , ranging from 1.3x to 3.5x; QuickScorer , ranging from 2.1x to 25.0x; VPred , ranging from 2.3x to 18.3x; and XGBoost , ranging from 2.6x to 42.5x.

Question-oriented text retrieval, aka natural language-based text retrieval, has been widely used in software engineering. Earlier work has concluded that questions with the same keywords but different interrogatives (such as how, what) should result in different answers. But what is the difference? How to identify the right answers to a question? In this paper, we propose to investigate the "answer style" of software questions with different interrogatives. Towards this end, we build classifiers in a software text repository and propose a re-ranking approach to refine search results. The classifiers are trained by over 16,000 answers from the StackOverflow forum. Each answer is labeled accurately by its question's explicit or implicit interrogatives. We have evaluated the performance of our classifiers and the refinement of our re-ranking approach in software text retrieval. Our approach results in 13.1% and 12.6% respectively improvement with respect to text retrieval criteria nDCG@1 and nDCG@10 compared to the baseline. We also apply our approach to FAQs of 7 open source projects and show 13.2% improvement with respect to nDCG@1. The results of our experiments suggest that our approach could find answers to FAQs more precisely.

In many software engineering tasks, question-oriented text retrieval is often used to help developers search for software artifacts. In this paper, we propose an interrogative-guided re-ranking approach for question-oriented software text retrieval. Since different interrogatives usually indicate users' different search focuses, we firstly label 9 kinds of question-answer pairs according to the common interrogatives. Then, we train document classifiers by using 1,826 questions along with 2,460 answers from StackOverflow, apply the classifiers to our document repository and present a re-ranking approach to improve the retrieval precision. In software document classification, our classifiers achieve the average precision, recall and F-measure of 56.2%, 90.9% and 69.4% respectively. Our re-ranking approach presents 9.6% improvement in [email protected] upon the baseline, and we also obtain 8.1% improvement in [email protected] when more candidates are included.