Quantitative Optimization-Intensive Approach, Miodrag Potkonjak, Jan M. Rabaey · 1999
Implementation platform selection is an important component of hardware--software codesign process which selects, for a given computation, the most suitable implementation platform. In this paper, we study the complementary component of hardware--software codesign, algorithm selection. Given a set of specifications for the targeted application, algorithm selection refers to choosing the most suitable completely specified computational structure for a given set of design goals and constraints, among several functionally equivalent alternatives. While implementation platform selection has been recently widely and vigorously studied, the algorithm selection problem has not been studied in computer-aided design domain until now.
Cited by 0
Mi-yeon Yu (813778), Sung Woo Lee (747615), Seon Ha Baek (498952) et al. · Figshare · 2017
Cited by 0Open access
Mustafa Mısır · 2022 IEEE Symposium Series on Computational Intelligence (SSCI) · 2022
The present study introduces algorithm selection on selection hyper-heuristics. Hyper-heuristics are known as problem-independent methods utilized to solve different instances from varying problem domains. In the literature, there has been effective hyper-heuristic designs providing a certain level of generality in problem solving. Still, the relevant existing research indicates that there is no single hyper-heuristic which performs always the best on different problem solving scenarios. Algorithm selection has been investigated essentially to address this issue, mainly for the problem-specific algorithms, by automatically identifying the (near) best algorithm(s) for each given problem instance. This paper performs algorithm selection on selection hyper-heuristics, for the first time, delivering cross-domain algorithm selection. For this purpose, a suite of problem-independent features is initially introduced. Then, algorithm selection is examined across 9 single-objective combinatorial optimization problems with 6 online selection hyper-heuristics. The experimental results carried out on these problems indicated that algorithm selection is effective for choosing hyper-heuristics while offering improved generality and robustness.
Cited by 5
Alex Gough (16734033), Alice Sitch (396379), Tom Marshall (154542) · Figshare · 2025
<div> Background C-reactive protein (CRP) is a biomarker of inflammation used in diagnosis of inflammatory diseases and to guide treatment decisions. Variation in within-individual measured CRP may affect its clinical utility but estimates of within-individual variation are based on limited data and so may not be accurate. Methods A retrospective cohort study was performed using data on CRP results and sociodemographic, lifestyle and comorbidity covariates extracted from the IQVIA Medical Research Database (IMRD) database using the DEXTER tool. A minimum of four measurements for each individual was the only inclusion criterion. CRP data were log-transformed for analysis. Within-individual measured variation was calculated as a coefficient of variation (CV) using a linear regression random effects model for the whole population and various subgroups. Results 472,811 participants were included in this study, making it the largest study of variation of CRP to date by a factor of approximately five. The overall coefficient of variation for CRP was 1.604 (95% CI 1.602 to 1.606). This is much higher than the median reported CV for CRP of previous studies which was 0.41. CV increased with patient median. Strengths and limitations The large number of participants and the real-world nature of the results are important strengths of this study. Weaknesses included the problem of accounting for confounding by indication, and the short half-life of CRP making it hard to distinguish between acute illness and physiological variation. Conclusions Estimated within-individual variation in this analysis of real-world data is very high and is higher than previously reported. Variation increases with patient median CRP, that is with more severe disease status. This has important implications for the diagnosis, monitoring and clinical decision-making for inflammatory disease. </div>
Cited by 0Open access
Daniel R. Zollinger (16713138), Elizabeth Rivers (18570199), Alexander Fine (9625713) et al. · Figshare · 2024
<p>A) ROC curve demonstrating the True Positive Rate and False Positive Rates at differing somatic probability thresholds. Black dashed line indicates x = y. The vertical orange dashed line corresponds to a FPR of 0.01. B) Flowchart for the inclusion of short variant alterations in patient specific design.</p> <p>(TIF)</p>
Cited by 0Open access
Daniel R. Zollinger (16713138), Elizabeth Rivers (18570199), Alexander Fine (9625713) et al. · Figshare · 2024
<p>A) Percentage of patients by number of monitorable alterations and by variant status. B) Number of monitorable alterations designed per patient from historical tissue CGP across tumor types based on Foundation Medicine disease classification (See <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0302129#pone.0302129.s004" target="_blank">S1 Table</a> for disease mapping). C) PPV of variant selection algorithm at the variant and sample-level. D) Frequency of samples with false positive (FP) variants. Error bars = 90% CI. Abbreviations: crc = colorectal cancer, nsclc = non-small cell lung cancer, cup = cancer of unknown primary (physician specified), gi- gastrointestinal, pns = paranasal sinus, cns = central nervous system, gist = gastrointestinal stromal tumor, FP = false positive, PPV = positive predictive value.</p>
Cited by 0Open access
Alex Gough (16734033), Tom Marshall (154542), Alice Sitch (396379) · Figshare · 2025
<div> Background HbA1c is a marker for diabetes mellitus that reflects average glucose concentrations over the previous eight to twelve weeks. It is used to aid in the diagnosis and management of diabetes. Variation in within-individual measured HbA1c may affect its clinical utility but estimates of this are based on limited data that is often not generalisable to real-world settings. Methods A retrospective cohort study was performed using data on HbA1c results and sociodemographic, lifestyle and comorbidity covariates extracted from the IQVIA Medical Research Database (IMRD) database using the DEXTER tool. A minimum of four measurements in the same individual was the only inclusion criterion. Within-individual measured variation was calculated as a coefficient of variation (CV) using a linear regression random effects model for the whole population and various subgroups. Results 587,023 participants were included in this study, making it the largest study of variation of HbA1c to date. The overall measured within-individual coefficient of variation (CV<sub>T</sub>) was 0.20 (95%CI 0.20 to 0.20). This is around three times higher than reported in a previous systematic review. CV<sub>T</sub> increased with patient mean HbA1c level. Strengths and limitations The large number of participants and the real-world nature of the results are important strengths of this study. Weaknesses included the problem of accounting for confounding by indication. Conclusions Estimated within-individual variation in this analysis of real-world data is very high and is higher than previously reported. Variation increases with patient mean HbA1c, that is with more severe disease status. This has important implications for the diagnosis, monitoring and clinical decision-making for diabetes. </div>
Cited by 0Open access
Dwi Kisworo Setyowireni (22656144), Roni Naning (22656147), Rina Triasih (812588) et al. · Figshare · 2025
<div> Background Bronchopulmonary dysplasia (BDP) is the most common chronic lung disease in infants. The study aimed to evaluate the outcome of infants diagnosed with BPD in Yogyakarta, Indonesia. Material and methods A retrospective cohort study was conducted by reviewing the medical records of infants with BPD born between January 2015 and December 2020 at Dr. Sardjito General Hospital in Yogyakarta, Indonesia. Surviving children were invited for clinical assessment and echocardiogram. A multivariate logistic regression analysis was used to identify predictors of mortality. Results Among 8,490 newborns hospitalized in the Perinatal ward, 100 (1.2%) developed BPD, with neonatal sepsis and congenital heart disease as the most prevalent comorbidities. Of the 85 infants with complete data, 41 (48.2%) died within 7 months. Multivariate analysis revealed that post-menstrual age (PMA) < 28 weeks and mechanical ventilation as independent predictors of mortality, with adjusted odds ratios (ORs) (95% CI) of 5.27 (1.36-20.43) and 0.26 (0.09-0.75), respectively. Conclusions Nearly half of the infants diagnosed with BPD died within the first seven months of life. PMA < 28 weeks was associated with increased mortality risk, while mechanical ventilation showed a protective effect against mortality in infants with BPD. </div>
Cited by 0Open access
Jaiberth Antonio Cardona-Arias (744735), Iván Darío Vélez (744736), Liliana López-Carvajal (744737) · Figshare · 2015
Cited by 0Open access
Simone Rubinacci (9647929), Olivier Delaneau (76754), Jonathan Marchini (21497) · Figshare · 2020
Cited by 0Open access
Braulio Josué Méndez-Sotelo (18405243), Mónica Delgado-Beltrán (18405246), Melissa Hernández-Durán (7862702) et al. · Figshare · 2024
<div><p>Background</p><p>Infections caused by <i>Stenotrophomonas maltophilia</i> and related species are increasing worldwide. Unfortunately, treatment options are limited, whereas the antimicrobial resistance is increasing.</p><p>Methods</p><p>We included clinical isolates identified as <i>S</i>. <i>maltophilia</i> by VITEK 2 Compact. Ceftazidime/avibactam, meropenem/vaborbactam, imipenem/relebactam, cefiderocol, quinolones, and tetracycline family members were evaluated by broth microdilution method and compared with first-line treatment drugs. Minimum inhibitory concentrations (MICs) were reported for all antibiotics. We sequenced the Whole Genome of cefiderocol resistant strains (CRSs) and annotated their genes associated with cefiderocol resistance (GACR). Presumptive phylogenetic identification employing the 16S marker was performed.</p><p>Results</p><p>One hundred and one clinical strains were evaluated, sulfamethoxazole and trimethoprim, levofloxacin and minocycline showed susceptibilities of 99.01%, 95.04% and 100% respectively. Ceftazidime was the antibiotic with the highest percentage of resistance in all samples (77.22%). Five strains were resistant to cefiderocol exhibiting MIC values ≥ 2 μg/mL (4.95%). The β-lactamase inhibitors meropenem/vaborbactam and imipenem/relebactam, failed to inhibit <i>S</i>. <i>maltophilia</i>, preserving both MIC<sub>50</sub> and MIC<sub>90</sub> ≥64 μg/mL. Ceftazidime/avibactam restored the activity of ceftazidime decreasing the MIC range. Tigecycline had the lowest MIC range, MIC<sub>50</sub> and MIC<sub>90.</sub> Phylogeny based on 16S rRNA allowed to identify to cefiderocol resistant strains as putative species clustered into <i>Stenotrophomonas maltophilia</i> complex (Smc). In these strains, we detected GARCs such as Mutiple Drug Resistance (MDR) efflux pumps, L1-type β-lactamases, iron transporters and type-1 fimbriae.</p><p>Conclusion</p><p>Antimicrobial resistance to first-line treatment is low. The <i>in vitro</i> activity of new β-lactamase inhibitors against <i>S</i>. <i>maltophilia</i> is poor, but avibactam may be a potential option. Cefiderocol could be considered as a potential new option for multidrug resistant infections. Tetracyclines had the best <i>in vitro</i> activity of all antibiotics evaluated.</p></div>
Cited by 0Open access
Simone Rubinacci (9647929), Olivier Delaneau (76754), Jonathan Marchini (21497) · Figshare · 2020
Cited by 0Open access
Charmaine Demanuele (86247), Peter Kirsch (178136), Christine Esslinger (386168) et al. · Figshare · 2015
<p>(<i>A</i>) Penalized cross validation error for task-related (blue) and non-task-related (red) ROIs. (<i>B</i>) Penalized prediction error curve for the left DLPFC, smoothed by a Gaussian kernel (cyan). A minimum (non-penalized) prediction error of 0.327 = 0.505/(1 + ξ<i>p</i>/log<i>n</i>) was obtained for the overall separation of the six cognitive task stages; <i>p</i> is the number of voxels for this ROI (left DLPFC), <i>n</i> is the number of time points, and ξ = 1. (<i>C</i>) The most predictive voxels (marked in red) selected from a combined ROI mask (blue regions) were distributed across the DLPFC, OFC, ACC and hippocampus.</p>
Cited by 0Open access
Erdem Kuş, Özgür Akgün, Nguyen Dang et al. · arXiv (Cornell University) · 2024
When solving decision and optimisation problems, many competing algorithms (model and solver choices) have complementary strengths. Typically, there is no single algorithm that works well for all instances of a problem. Automated algorithm selection has been shown to work very well for choosing a suitable algorithm for a given instance. However, the cost of training can be prohibitively large due to running candidate algorithms on a representative set of training instances. In this work, we explore reducing this cost by choosing a subset of the training instances on which to train. We approach this problem in three ways: using active learning to decide based on prediction uncertainty, augmenting the algorithm predictors with a timeout predictor, and collecting training data using a progressively increasing timeout. We evaluate combinations of these approaches on six datasets from ASLib and present the reduction in labelling cost achieved by each option.
Cited by 0Open access
Jianjun Sun (54027), Li Liu (75607), Jiayin Shen (775304) et al. · Figshare · 2015
Cited by 0Open access
Mahmoud Zeydabadinezhad (12289570), Charles C. Horn (7517456), Babak Mahmoudi (354650) · Figshare · 2023
<div><p>Vagus nerve stimulation (VNS) is a potential treatment option for gastrointestinal (GI) diseases. The present study aimed to understand the physiological effects of VNS on gastrointestinal (GI) function, which is crucial for developing more effective adaptive closed-loop VNS therapies for GI diseases. Electrogastrography (EGG), which measures gastric electrical activities (GEAs) as a proxy to quantify GI functions, was employed in our investigation. We introduced a recording schema that allowed us to simultaneously induce electrical VNS and record EGG. While this setup created a unique model for studying the effects of VNS on the GI function and provided an excellent testbed for designing advanced neuromodulation therapies, the resulting data was noisy, heterogeneous, and required specialized analysis tools. The current study aimed at formulating a systematic and interpretable approach to quantify the physiological effects of electrical VNS on GEAs in ferrets by using signal processing and machine learning techniques. Our analysis pipeline included pre-processing steps, feature extraction from both time and frequency domains, a voting algorithm for selecting features, and model training and validation. Our results indicated that the electrophysiological changes induced by VNS were optimally characterized by a distinct set of features for each classification scenario. Additionally, our findings demonstrated that the process of feature selection enhanced classification performance and facilitated representation learning.</p></div>
Cited by 0Open access
Sung Woo Lee (747615), Sejoong Kim (286893), Ki Young Na (299530) et al. · Figshare · 2016
Cited by 0Open access
Marc W. Chalé · 2020
Cited by 0
Fuyong Zhang, Ying Ma · 2016
Run-time malware detection strategies are efficient and robust, which get more and more attention. In this paper, we use I/O Request Package (IRP) sequences for malware detection. N-gram will be used to analyze IRP sequences for feature extraction. Integrated Negative Selection Algorithm (NSA) and Positive Selection Algorithm (PSA), through a selection of n-gram sequences which only exist in malware IRP sequences, we have more than 96% true positive rate and 0% false positive rate.
Cited by 7
Parth Gupta, G. P. Bhattacharjee · BIT Numerical Mathematics · 1984
Cited by 1