46 research outputs found
A study on relation extraction of historical figures based on bibliographic description
No full textFigure relation extraction is an important and hard field in information extraction. In this paper, aiming to improve the performance for relation extraction of historical figures, we propose a novel method based on bibliographic description. In the proposed method, by analyzing the species and co-occurrence relation of responsibility in a bibliographic record, we combine diverse person responsibility, person name and time as features, whose values are the quantity of the species clustering concerned, to build a Decision Tree model. Accordingly, relation extraction of historical figures is performed through the model. It is experimentally shown that on average, 83.3% and 83.0% in precision and recall rate are achieved respectively without more linguistic knowledge and complex classifiers. ? 2011 IEEE.EI
Systematic Analysis of Aberrant Biochemical Networks and Potential Drug Vulnerabilities Induced by Tumor Suppressor Loss in Malignant Pleural Mesothelioma.
No full textBackground: Malignant pleural mesothelioma (MPM) is driven by the inactivation of tumor suppressor genes (TSGs). An unmet need in the field is the translation of the genomic landscape into effective TSG-specific therapies. Methods: We correlated genomes against transcriptomes of patients' MPM tumors, by weighted gene co-expression network analysis (WGCNA). The identified aberrant biochemical networks and potential drug targets induced by tumor suppressor loss were validated by integrative data analysis and functional interrogation. Results: CDKN2A/2B loss activates G2/M checkpoint and PI3K/AKT, prioritizing a co-targeting strategy for CDKN2A/2B-null MPM. CDKN2A deficiency significantly co-occurs with deletions of anti-viral type I interferon (IFN-I) genes and BAP1 mutations, that enriches the IFN-I signature, stratifying a unique subset, with deficient IFN-I, but proficient BAP1 for oncolytic viral immunotherapies. Aberrant p53 attenuates differentiation and SETD2 loss acquires the dependency on EGFRs, highlighting the potential of differentiation therapy and pan-EGFR inhibitors for these subpopulations, respectively. LATS2 deficiency is linked with dysregulated immunoregulation, suggesting a rationale for immune checkpoint blockade. Finally, multiple lines of evidence support Dasatinib as a promising therapeutic for LATS2-mutant MPM. Conclusions: Systematic identification of abnormal cellular processes and potential drug vulnerabilities specified by TSG alterations provide a framework for precision oncology in MPM
Null broadening for multistage nested Wiener filter on the joint space-time GPS receiver
No full text
Analysis on urban construction land changes and influential factors of Yangtze River Delta urban agglomerations
No full textNan Song Shuzhou gong du yi jian zheng li yu yan jiu Zheng li yu yan jiu
No full textBen shu shi dui wo she yu 1990 nian chu ban de ying yin ben "Song ren yi jian" de zheng li yu yan jiu. Gai shu shi liao feng fu, she ji zheng zhi, jing ji, jun shi yi ji Song dai gong wen ge shi he shu yi, shi zhen gui de shi wu wen xian. "Song ren yi jian" fan ying le bu shao nan Song shi qi di fang guan fu ji gou she zhi he xing zheng guan li zhi du, bu chong le xu duo Song dai cai zheng, shui shou zheng ce de xin cai liao, bao liu le nan Song chu nian Jiang Huai di qu zhan bei zhuang kuang de xi jie cai liao, hai shi Song dai wen shu zhi du de zhong yao cai liao, you zhu yu Song shi yan jiu de shen ru. Zheng li ben yan jin gui fan, ge shi ming xi, ji ben huan yuan le Song jian tu ban de ben lai mian mao, ju you jiao qiang de can kao jia zh
Duo chong zu xue shu ju fen xi
No full textPh.D.High-throughput omics technologies have been widely adopted in current biomedical research. Scientists take advantage of extensive information in omics data to explore the molecular mechanisms underlying biological process and to find novel biomarkers or therapeutic targets for various diseases. Recently, researchers introduce the “multi-omics” strategy, where we generate and combine multiple omics data either of the same or different types to obtain a more holistic view of the complicated system. The rapid development and popularization of multi-omics design are followed by an urgent requirement of rigorous methods for integrative analysis. In this thesis, we propose two novel approaches to omics data integration.This thesis is mainly composed of two parts. In the first part, we propose a statistical method called “XMotif” for jointly analyzing different types of epigenetic data across dozens of cell types. We want to answer the question of how the combinatorial epigenetic signals vary along the genome and across diverse cell types. We apply XMotif to DNA methylation and chromatin accessibility data from 69 ENCODE cell lines and acquire 90 combinatorial epigenetic patterns. In analogy to sequence motifs, these dynamic epigenetic patterns annotate the genome with parsimonious representations. Accordingly, we name them as epigenome motifs (emotifs). Genomic regions coded by the same emotif share similar biological functions.In the second part, we introduce a meta-analysis model named “PanDM” that integrates methylome data from multiple cancer types for pan-cancer differential DNA methylation analysis. The PanDM model takes summary statistics from separate analyses as input and carries out methylation site clustering, differential methylation detection, as well as pan-cancer pattern discovery simultaneously. We demonstrate the performance of our model using simulation data. We also apply PanDM to the real data collected from The Cancer Genome Atlas (TCGA) and discover novel pan-cancer differential methylation patterns. The PanDM is implemented in an R package, which can serve as a useful resource to experimentalists who work on cancer epigenomics.高通量組學技術在生物醫學研究中得到了廣泛的應用。科學家們利用組學數據中的大量信息來探索生物學過程背後的分子機制，並為各種疾病尋找新的生物標誌物或治療靶點。近年來，研究人員引入了“多組學”策略，在該策略中，我們生成並組合不同類型或相同類型的多個組學數據，以獲得對復雜系統更全面的看法。隨著多組學設計的迅速發展和普及，對集成分析方法的嚴格要求也應運而生。本文提出了兩種新的組學數據集成方法。本文主要由两部分組成。在第一部分中，我們提出了一種名為“XMotif”的統計方法，用於聯合分析數十個細胞類型的不同類型的表觀遺傳數據。我們想要回答的問題是組合表觀遺傳信號是如何隨基因組和不同細胞類型而變化的。我們將XMotif應用於69個ENCODE細胞系的DNA甲基化和染色質可及性數據，獲得90個組合表觀遺傳模式。與序列模體類似，這些動態的表觀遺傳模式用簡潔的表示來標註基因組。因此，我們將它們命名為表觀基因組模體（emotifs）。用同一個表觀基因組模體編碼的基因組區域具有相同的生物學功能。在第二部分中，我們引入了一個名為“PanDM”的元分析模型，該模型整合了來自多種癌癥類型的甲基化組數據，來進行泛癌癥差異DNA甲基化分析。PanDM模型以獨立分析的匯總統計數據作為輸入，同時進行甲基化位點聚類、差異甲基化檢測、泛癌模式發現。我們使用仿真數據演示了模型的性能。我們也將PanDM應用於從癌癥基因組圖譜（TCGA）收集到的真實數據上，並發現了新的泛癌癥的甲基化模式。我們將PanDM制作成了一個R包，都可以作為研究癌癥表觀基因組學的實驗家們的有用資源。Shi, Mai.Thesis Ph.D. Chinese University of Hong Kong 2018.Includes bibliographical references (leaves 100-110).Abstracts also in Chinese.Title from PDF title page (viewed on …).Shi, Mai
Voltage-mode variable frequency control for single-inductor dual-output buck converter with fast transient response
No full textLearning and control in trustworthy and responsible artificial intelligence for cyber-physical systems
No full textWith breakthroughs in technology and excellence in robustness, artificial intelligence (AI) has gradually reshaped the way of information interaction and has been challenging traditional cyber-physical systems (CPS) research. Nourished by enormous data, AI and data-driven machine learning have been emerging with CPS ideas for better capability and adaptability in automation and autonomy. Yet, concerns about the emerging ideas have been raised on dependability and accountability. The data-driven models have been proven to over-reliance on training data, which exposes unreliable generalization and is sensitive to data disturbances. It is true that a larger and broader training dataset could tend to a better model, however, gathering such a dataset is a time-consuming, labor-intensive, and cost-demanding progress with threats to individual privacy. Therefore, it is inevitable to find ethical and robust alternative solutions, among which an interest is in trustworthy and responsible AI. This dissertation explores the learning and control in trustworthy and responsible AI for CPS from a control perspective. In particular, this dissertation intends to improve the performance of data-driven learning on non-cooperative facial-credentialed security authentication within complex backgrounds.
Limited by the physical and cyber components, the quality of visual data within CPS is generally worse than that in datasets. Typically, the collection of datasets for facial-credentialed security authentication tends to address variation in pose, size, illumination, and occlusion but not in facial quality (scales). However, facial quality is an inevitable factor affecting non-cooperative facial-credentialed security authentication. Feature Super-Resolution Face Recognition Net (FSRFR-Net) is designed to balance the quality-effect on facial-credentialed recognition. It utilizes the trained data-driven models, which are convolutional neural networks trained on high-quality faces, and improves their performance on low-quality faces with a minimum change, which is to insert a multi-branch feature super resolution module, in the structure. Experiments on different datasets present consistency in performance improvement on low-quality recognition while preserving high-quality recognition performance. The performance improvement difference among datasets demonstrates the high correlation between the larger face region and better recognition performance.
Following that observation, an anchor-free duo-agent reinforcement learning model is proposed for multi-scale (multi-quality) face detection. It measures the image quality caused influence from the weighted combination of channel-wise feature maps. The design of duo-agent allows the freedom in the pixel-size (resolution) of faces while preserving computational simplicity. The detection goal is to maximize the facial region while minimize the background region. In addition, the duo-agent reinforcement learning design transforms the face detection problem into a classic global vs. local extrema regression problem in the 3D feature space. Consequentially, an intention is to find a stable and convergence solution to the regression problem.
In order to solve the stochastic discrete-time model-free duo-agent regression problem, the convergence efficiency and stability of reinforcement learning approaches in CPS have been explored from the control theory perspective. A mild condition has been proven applicability to policy iteration approach of zero-sum differential games and H∞ robust optimal control. A Simultaneous Online Tracking and Planning (SOTP) Navigation system of autonomous tracked vehicles for leader-follower formation is proposed to examine the optimal strategies obtained through policy iteration approach of multi-agent linear-quadratic optimal control problems. The SOTP framework is analyzed and designed through a skid-steering dynamic model with optimization of low-level controls (agents) of autonomous tracked vehicles for the leader-follower formation. Simulation results have shown the efficacy and robustness of the SOTP framework while the freedom of trajectory curvature and formation posture are reserved without the virtual followers/leaders.Ph.D.Includes bibliographical reference
