Yajun Mei
Yajun Mei
Scroll
Professor of Biostatistics
-
Professional overview
-
Yajun Mei is a Professor of Biostatistics at NYU/GPH, starting from July 1, 2024. He received the B.S. degree in Mathematics from Peking University, Beijing, China, in 1996, and the Ph.D. degree in Mathematics with a minor in Electrical Engineering from the California Institute of Technology, Pasadena, CA, USA, in 2003. He was a Postdoc in Biostatistics in the renowned Fred Hutch Cancer Center in Seattle, WA during 2003 and 2005. Prior to joining NYU, Dr. Mei was an Assistant/Associate/Full Professor in H. Milton Stewart School of Industrial and Systems Engineering at the Georgia Institute of Technology, Atlanta, GA for 18 years from 2006 to 2024, and had been a co-director of Biostatistics, Epidemiology, and Study Design (BERD) of Georgia CTSA since 2018.
Dr. Mei’s research interests are statistics, machine learning, and data science, and their applications in biomedical science and public health, particularly, streaming data analysis, sequential decision/design, change-point problems, precision/personalized medicine, hot-spots detection for infectious diseases, longitudinal data analysis, bioinformatics, and clinical trials. His work has received several recognitions including Abraham Wald Prizes in Sequential Analysis in both 2009 and 2024, NSF CAREER Award in 2010, an elected Fellow of American Statistical Association (ASA) in 2023, and multiple best paper awards.
-
Education
-
BS, Mathematics, Peking UniversityPhD, Mathematics, California Institute of Technology
-
Honors and awards
-
Fellow of American Statistical Association (2023)Star Research Achievement Award, 2021 Virtual Critical Care Congress (2021)Best Paper Competition Award, Quality, Statistics & Reliability of INFORMS (2020)Bronze Snapshot Award, Society of Critical Care Medicine (2019)NSF Career AwardThank a Teacher Certificate, Center for Teaching and Learning (2011201220162020202120222023)Abraham Wald Prize (2009)Best Paper Award, 11th International Conference on Information Fusion (2008)New Researcher Fellow, Statistical and Applied Mathematical Sciences Institute (2005)Fred Hutchinson SPAC Travel Award to attend 2005 Joint Statistical Meetings, Minneapolis, MN (2005)Travel Award to 8th New Researchers Conference, Minneapolis, MN (2005)Travel Award to IEEE International Symposium on Information Theory, Chicago, IL (2004)Travel Award to IPAM workshop on inverse problem, UCLA, Los Angeles, CA (2003)Fred Hutchinson SPAC Course Scholarship (2003)Travel Award to the SAMSI workshop on inverse problem, Research Triangular Park, NC (2002)
-
Publications
Publications
Hot-spots detection in count data by Poisson assisted smooth sparse tensor decomposition
AbstractZhao, Y., Huo, X., & Mei, Y. (n.d.).Publication year
2023Journal title
Journal of Applied StatisticsVolume
50Issue
14Page(s)
2999-3029AbstractCount data occur widely in many bio-surveillance and healthcare applications, e.g. the numbers of new patients of different types of infectious diseases from different cities/counties/states repeatedly over time, say, daily/weekly/monthly. For this type of count data, one important task is the quick detection and localization of hot-spots in terms of unusual infectious rates so that we can respond appropriately. In this paper, we develop a method called Poisson assisted Smooth Sparse Tensor Decomposition (PoSSTenD), which not only detect when hot-spots occur but also localize where hot-spots occur. The main idea of our proposed PoSSTenD method is articulated as follows. First, we represent the observed count data as a three-dimensional tensor including (1) a spatial dimension for location patterns, e.g. different cities/countries/states; (2) a temporal domain for time patterns, e.g. daily/weekly/monthly; (3) a categorical dimension for different types of data sources, e.g. different types of diseases. Second, we fit this tensor into a Poisson regression model, and then we further decompose the infectious rate into two components: smooth global trend and local hot-spots. Third, we detect when hot-spots occur by building a cumulative sum (CUSUM) control chart and localize where hot-spots occur by their LASSO-type sparse estimation. The usefulness of our proposed methodology is validated through numerical simulation studies and a real-world dataset, which records the annual number of 10 different infectious diseases from 1993 to 2018 for 49 mainland states in the United States.Pivotal Estimation of Linear Discriminant Analysis in High Dimensions
AbstractFang, E. X., Mei, Y., Shi, Y., Xu, Q., & Zhao, T. (n.d.).Publication year
2023Journal title
Journal of Machine Learning ResearchVolume
24AbstractWe consider the linear discriminant analysis problem in the high-dimensional settings. In this work, we propose PANDA(PivotAl liNear Discriminant Analysis), a tuning-insensitive method in the sense that it requires very little effort to tune the parameters. Moreover, we prove that PANDA achieves the optimal convergence rate in terms of both the estimation error and misclassification rate. Our theoretical results are backed up by thorough numerical studies using both simulated and real datasets. In comparison with the existing methods, we observe that our proposed PANDA yields equal or better performance, and requires substantially less effort in parameter tuning.Repetitive Low-level Blast Exposure and Neurocognitive Effects in Army Ranger Mortarmen
AbstractWoodall, J. L., Sak, J. A., Cowdrick, K. R., Muñoz, B. M., McElrath, J. H., Trimpe, G. R., Mei, Y., Myhre, R. L., Rains, J. K., & Hutchinson, C. R. (n.d.).Publication year
2023Journal title
Military MedicineVolume
188Issue
3-4Page(s)
E771-E779AbstractIntroduction: Occupational exposure to repetitive, low-level blasts in military training and combat has been tied to subconcussive injury and poor health outcomes for service members. Most low-level blast studies to date have focused on explosive breaching and firing heavy weapon systems; however, there is limited research on the repetitive blast exposure and physiological effects that mortarmen experience when firing mortar weapon systems. Motivated by anecdotal symptoms of mortarmen, the purpose of this paper is to characterize this exposure and its resulting neurocognitive effects in order to provide preliminary findings and actionable recommendations to safeguard the health of mortarmen. Materials and Methods: In collaboration with the U.S. Army Rangers at Fort Benning, blast exposure, symptoms, and pupillary light reflex were measured during 3 days of firing 81 mm and 120 mm mortars in training. Blast exposure analysis included the examination of the blast overpressure (BOP) and cumulative exposure by mortarman position, as well as comparison to the 4 psi safety threshold. Pupillary light reflex responses were analyzed with linear mixed effects modeling. All neurocognitive results were compared between mortarmen (n=11) and controls (n=4) and cross-compared with blast exposure and blast history. Results: Nearly 500 rounds were fired during the study, resulting in a high cumulative blast exposure for all mortarmen. While two mortarmen had average BOPs exceeding the 4 psi safety limit (Fig. 2), there was a high prevalence of mTBIlike symptoms among all mortarmen, with over 70% experiencing headaches, ringing in the ears, forgetfulness/poor memory, and taking longer to think during the training week (n≥8/11). Mortarmen also had smaller and slower pupillary light reflex responses relative to controls, with significantly slower dilation velocity (PRobust High-Dimensional Linear Discriminant Analysis under Training Data Contamination
AbstractShi, Y., Deshmukh, A., Mei, Y., & Veeravalli, V. (n.d.).Publication year
2023Page(s)
2099-2104AbstractThe problem of robust Sparse Linear Discriminant Analysis (LDA) in high-dimensions is studied, in which a fraction of the training data may be corrupted by an adversary. A computationally efficient algorithm is proposed by adapting robust mean estimation along with a calibration framework for LDA. Theoretical properties of the proposed algorithm are established for both the estimation error of the optimal projection vector and the mis-classification rate. Results from extensive numerical studies on both synthetic and real datasets are reported to show the usefulness of our algorithm.Surge Capacity in the COVID-19 Era : a Natural Experiment of Neurocritical Care in General Critical Care
AbstractPhilips, S., Shi, Y., Coopersmith, C. M., Samuels, O. B., Pimentel-Farias, C., Mei, Y., Sadan, O., & Akbik, F. (n.d.).Publication year
2023Journal title
Neurocritical CareVolume
38Issue
2Page(s)
320-325AbstractBackground: COVID-19 surges led to significant challenges in ensuring critical care capacity. In response, some centers leveraged neurocritical care (NCC) capacity as part of the surge response, with neurointensivists providing general critical care for patients with COVID-19 without neurologic illness. The relative outcomes of NCC critical care management of patients with COVID-19 remain unclear and may help guide further surge planning and provide broader insights into general critical care provided in NCC units. Methods: We performed an observational cohort study of all patients requiring critical care for COVID-19 across four hospitals within the Emory Healthcare system during the first three surges. Patients were categorized on the basis of admission to intensive care units (ICUs) staffed by general intensivists or neurointensivists. Patients with primary neurological diagnoses were excluded. Baseline demographics, clinical complications, and outcomes were compared between groups using univariable and propensity score matching statistics. Results: A total of 1141 patients with a primary diagnosis of COVID-19 required ICU admission. ICUs were staffed by general intensivists (n = 1071) or neurointensivists (n = 70). Baseline demographics and presentation characteristics were similar between groups, except for patients admitted to neurointensivist-staffed ICUs being younger (59 vs. 65, p = 0.027) and having a higher PaO2/FiO2 ratio (153 vs. 120, p = 0.002). After propensity score matching, there was no correlation between ICU staffing and the use of mechanical ventilation, renal replacement therapy, and vasopressors. The rates of in-hospital mortality and hospice disposition were similar in neurointensivist-staffed COVID-19 units (odds ratio 0.9, 95% confidence interval 0.31–2.64, p = 0.842). Conclusions: COVID-19 surges precipitated a natural experiment in which neurology-trained neurointensivists provided critical care in a comparable context to general intensivists treating the same disease. Neurology-trained neurointensivists delivered comparable outcomes to those of general ICUs during COVID-19 surges. These results further support the role of NCC in meeting general critical care needs of neurocritically ill patients and as a viable surge resource in general critical care.Active Quickest Detection When Monitoring Multi-streams with Two Affected Streams
AbstractAbstractWe study the multi-stream quickest detection problem under the active learning setup, It is assumed that there are p local streams in a system and s ≤ p unknown local streams are affected by an undesired event at some unknown time, but one is only able to take observations from r of these p local streams at each time instant. The objective is how to adaptively sample from these p local streams and how to use the observed data to raise a correct global alarm as quickly as possible. In this paper, we develop the first asymptotic optimality theory in the active quickest detection literature for the case when s = r = 2. To be more concrete, we propose to combine three ideas to develop efficient active quickest detection algorithms: (1) win-stay, lose-switch sampling strategy; (2) local CUSUM statistics for local monitoring; and (3) the SUM-Shrinkage technique to fuse local statistics into a global decision. We show that our proposed algorithms are asymptotically optimal in the sense of minimizing detection delay up to the second order subject to the false alarm constraint. Numerical studies are conducted to validate our theoretical results.Adaptive Partially Observed Sequential Change Detection and Isolation
AbstractZhao, X., Hu, J., Mei, Y., & Yan1, H. (n.d.).Publication year
2022Journal title
TechnometricsVolume
64Issue
4Page(s)
502-512AbstractHigh-dimensional data has become popular due to the easy accessibility of sensors in modern industrial applications. However, one specific challenge is that it is often not easy to obtain complete measurements due to limited sensing powers and resource constraints. Furthermore, distinct failure patterns may exist in the systems, and it is necessary to identify the true failure pattern. This work focuses on the online adaptive monitoring of high-dimensional data in resource-constrained environments with multiple potential failure modes. To achieve this, we propose to apply the Shiryaev–Roberts procedure on the failure mode level and use the multi-arm bandit to balance the exploration and exploitation. We further discuss the theoretical property of the proposed algorithm to show that the proposed method can correctly isolate the failure mode. Finally, extensive simulations and two case studies demonstrate that the change point detection performance and the failure mode isolation accuracy can be greatly improved.Asymptotic Theory of `1-Regularized PDE Identification from a Single Noisy Trajectory
AbstractHe, Y., Suh, N., Huo, X., Kang, S. H., & Mei, Y. (n.d.).Publication year
2022Journal title
SIAM-ASA Journal on Uncertainty QuantificationVolume
10Issue
3Page(s)
1012-1036AbstractWe provide a formal theoretical analysis on the PDE identification via the `1-regularized pseudo least square method from the statistical point of view. In this article, we assume that the differential equation governing the dynamic system can be represented as a linear combination of various linear and nonlinear differential terms. Under noisy observations, we employ local-polynomial fitting for estimating state variables and apply the `1 penalty for model selection. Our theory proves that the classical mutual incoherence condition on the feature matrix F and the β∗min-condition for the ground-truth signal β∗ are sufficient for the signed-support recovery of the `1-PsLS method. We run numerical experiments on two popular PDE models, the viscous Burgers and the Korteweg-de Vries (KdV) equations, and the results from the experiments corroborate our theoretical predictions.Cannabis Use Is Not Associated with Aneurysmal Subarachnoid Hemorrhage Complications or Outcomes
AbstractAkbik, F., Konan, H. D., Williams, K. P., Ermias, L. M., Shi, Y., Takieddin, O., Grossberg, J. A., Howard, B. M., Tong, F., Cawley, C. M., Mei, Y., Samuels, O. B., & Sadan, O. (n.d.).Publication year
2022Journal title
StrokeVolume
53Issue
8Page(s)
E375-E376Abstract~Does intrathecal nicardipine for cerebral vasospasm following subarachnoid hemorrhage correlate with reduced delayed cerebral ischemia? A retrospective propensity score-based analysis
AbstractSadan, O., Waddel, H., Moore, R., Feng, C., Feng, C., Mei, Y., Pearce, D. G., Kraft, J., Pimentel, C., Mathew, S., Akbik, F., Ameli, P., Taylor, A., Danyluk, L., Martin, K. S., Garner, K., Kolenda, J., Pujari, A., Asbury, W., … Samuels, O. (n.d.).Publication year
2022Journal title
Journal of NeurosurgeryVolume
136Issue
1Page(s)
115-124AbstractOBJECTIVE Cerebral vasospasm and delayed cerebral ischemia (DCI) contribute to poor outcome following subarachnoid hemorrhage (SAH). With the paucity of effective treatments, the authors describe their experience with intrathecal (IT) nicardipine for this indication. METHODS Patients admitted to the Emory University Hospital neuroscience ICU between 2012 and 2017 with nontraumatic SAH, either aneurysmal or idiopathic, were included in the analysis. Using a propensity-score model, this patient cohort was compared to patients in the Subarachnoid Hemorrhage International Trialists (SAHIT) repository who did not receive IT nicardipine. The primary outcome was DCI. Secondary outcomes were long-term functional outcome and adverse events. RESULTS The analysis included 1351 patients, 422 of whom were diagnosed with cerebral vasospasm and treated with IT nicardipine. When compared with patients with no vasospasm (n = 859), the treated group was significantly younger (mean age 51.1 ± 12.4 years vs 56.7 ± 14.1 years, p < 0.001), had a higher World Federation of Neurosurgical Societies score and modified Fisher grade, and were more likely to undergo clipping of the ruptured aneurysm as compared to endovascular treatment (30.3% vs 11.3%, p < 0.001). Treatment with IT nicardipine decreased the daily mean transcranial Doppler velocities in 77.3% of the treated patients. When compared to patients not receiving IT nicardipine, treatment was not associated with an increased rate of bacterial ventriculitis (3.1% vs 2.7%, p > 0.1), yet higher rates of ventriculoperitoneal shunting were noted (19.9% vs 8.8%, p < 0.01). In a propensity score comparison to the SAHIT database, the odds ratio (OR) to develop DCI with IT nicardipine treatment was 0.61 (95% confidence interval [CI] 0.44-0.84), and the OR to have a favorable functional outcome (modified Rankin Scale score ≤ 2) was 2.17 (95% CI 1.61-2.91). CONCLUSIONS IT nicardipine was associated with improved outcome and reduced DCI compared with propensitymatched controls. There was an increased need for permanent CSF diversion but no other safety issues. These data should be considered when selecting medications and treatments to study in future randomized controlled clinical trials for SAH.Efficient Sequential UCB-based Hungarian Algorithm for Assignment Problems
AbstractAbstractThe assignment problem has many real-world applications such as allocations of agents and tasks for optimal utility gain. While it has been well-studied in the optimization literature when the underlying utility between every pair of agent and task is known, research is limited when the utilities are unknown and need to be learned from data on the fly. In this work, motivated by the mentor-mentee matching application in U.S. universities, we develop an efficient sequential assignment algorithm, with the objective of nearly maximizing the overall utility simultaneously for each time. Our proposed algorithm is to use stochastic binary bandit feedback to estimate the unknown utilities through the logistic regression, and then to combine the Upper Confidence Bound (UCB) method in the multi-armed bandit problem with the Hungarian algorithm in the assignment problem. We derive the theoretical bounds of our algorithm for both the estimation error and the total regret, and numerical studies are conducted to illustrate the usefulness of our algorithm.Implicit Regularization Properties of Variance Reduced Stochastic Mirror Descent
AbstractLuo, Y., Huo, X., & Mei, Y. (n.d.).Publication year
2022Page(s)
696-701AbstractIn machine learning and statistical data analysis, we often run into objective function that is a summation: the number of terms in the summation possibly is equal to the sample size, which can be enormous. In such a setting, the stochastic mirror descent (SMD) algorithm is a numerically efficient method - each iteration involving a very small subset of the data. The variance reduction version of SMD (VRSMD) can further improve SMD by inducing faster convergence. On the other hand, algorithms such as gradient descent and stochastic gradient descent have the implicit regularization property that leads to better performance in terms of the generalization errors. Little is known on whether such a property holds for VRSMD. We prove here that the discrete VRSMD estimator sequence converges to the minimum mirror interpolant in the linear regression. This establishes the implicit regularization property for VRSMD. As an application of the above result, we derive a model estimation accuracy result in the setting when the true model is sparse. We use numerical examples to illustrate the empirical power of VRSMD.Predicting the rheology of limestone calcined clay cements (LC3) : Linking composition and hydration kinetics to yield stress through Machine Learning
AbstractCanbek, O., Xu, Q., Mei, Y., Washburn, N. R., & Kurtis, K. E. (n.d.).Publication year
2022Journal title
Cement and Concrete ResearchVolume
160AbstractThe physicochemical characteristics of calcined clay influence yield stress of limestone calcined clay cements (LC3), but the independent influences the clay's physical and chemical characteristics as well as the effect of other variables on LC3 rheology are less well-understood. Further, a relationship between LC3 hydration kinetics and yield stress – important for informing mixture design – has not yet been established. Here, rheological properties were determined in pastes with varying water-to-solid ratio (w/s), constituent mass ratios (PC:metakaolin:limestone), limestone particle size and gypsum content. From these data, an ML model developed allowed the independent examination of the different mechanisms by which metakaolin fraction influences yield stress of LC3, identifying four predictors – packing index, Al2O3/SO3, total particle density and metakaolin fraction relative to limestone (MK/LS) – most significant for predicting LC3 yield stress. A methodology based on kernel smoothing also identified hydration kinetics parameters best correlated with yield stress.Private Sequential Hypothesis Testing for Statisticians : Privacy, Error Rates, and Sample Size
AbstractZhang, W., Mei, Y., & Cummings, R. (n.d.).Publication year
2022Journal title
Proceedings of Machine Learning ResearchVolume
151Page(s)
11356-11373AbstractThe sequential hypothesis testing problem is a class of statistical analyses where the sample size is not fixed in advance. Instead, the decision-process takes in new observations sequentially to make real-time decisions for testing an alternative hypothesis against a null hypothesis until some stopping criterion is satisfied. In many common applications of sequential hypothesis testing, the data can be highly sensitive and may require privacy protection; for example, sequential hypothesis testing is used in clinical trials, where doctors sequentially collect data from patients and must determine when to stop recruiting patients and whether the treatment is effective. The field of differential privacy has been developed to offer data analysis tools with strong privacy guarantees, and has been commonly applied to machine learning and statistical tasks. In this work, we study the sequential hypothesis testing problem under a slight variant of differential privacy, known as Renyi differential privacy. We present a new private algorithm based on Wald's Sequential Probability Ratio Test (SPRT) that also gives strong theoretical privacy guarantees. We provide theoretical analysis on statistical performance measured by Type I and Type II error as well as the expected sample size. We also empirically validate our theoretical results on several synthetic databases, showing that our algorithms also perform well in practice. Unlike previous work in private hypothesis testing that focused only on the classical fixed sample setting, our results in the sequential setting allow a conclusion to be reached much earlier, and thus saving the cost of collecting additional samples.Rapid detection of hot-spots via tensor decomposition with applications to crime rate data
AbstractZhao, Y., Yan, H., Holte, S., & Mei, Y. (n.d.).Publication year
2022Journal title
Journal of Applied StatisticsVolume
49Issue
7Page(s)
1636-1662AbstractIn many real-world applications of monitoring multivariate spatio-temporal data that are non-stationary over time, one is often interested in detecting hot-spots with spatial sparsity and temporal consistency, instead of detecting system-wise changes as in traditional statistical process control (SPC) literature. In this paper, we propose an efficient method to detect hot-spots through tensor decomposition, and our method has three steps. First, we fit the observed data into a Smooth Sparse Decomposition Tensor (SSD-Tensor) model that serves as a dimension reduction and de-noising technique: it is an additive model decomposing the original data into: smooth but non-stationary global mean, sparse local anomalies, and random noises. Next, we estimate model parameters by the penalized framework that includes Least Absolute Shrinkage and Selection Operator (LASSO) and fused LASSO penalty. An efficient recursive optimization algorithm is developed based on Fast Iterative Shrinkage Thresholding Algorithm (FISTA). Finally, we apply a Cumulative Sum (CUSUM) Control Chart to monitor model residuals after removing global means, which helps to detect when and where hot-spots occur. To demonstrate the usefulness of our proposed SSD-Tensor method, we compare it with several other methods including scan statistics, LASSO-based, PCA-based, T2-based control chart in extensive numerical simulation studies and a real crime rate dataset.Robust change detection for large-scale data streams
AbstractZhang, R., Mei, Y., & Shi, J. (n.d.).Publication year
2022Journal title
Sequential AnalysisVolume
41Issue
1Page(s)
1-19AbstractRobust change point detection for large-scale data streams has many real-world applications in industrial quality control, signal detection, and biosurveillance. Unfortunately, it is highly nontrivial to develop efficient schemes due to three challenges: (1) the unknown sparse subset of affected data streams, (2) the unexpected outliers, and (3) computational scalability for real-time monitoring and detection. In this article, we develop a family of efficient real-time robust detection schemes for monitoring large-scale independent data streams. For each data stream, we propose to construct a new local robust detection statistic called the (Formula presented.) -CUSUM (cumulative sum) statistic that can reduce the effect of outliers by using the Box-Cox transformation of the likelihood function. Then the global scheme will raise an alarm based upon the sum of the shrinkage transformation of these local (Formula presented.) -CUSUM statistics to filter out unaffected data streams. In addition, we propose a new concept called false alarm breakdown point to measure the robustness of online monitoring schemes and propose a worst-case detection efficiency score to measure the detection efficiency when the data contain outliers. We then characterize the breakdown point and the efficiency score of our proposed schemes. Asymptotic analysis and numerical simulations are conducted to illustrate the robustness and efficiency of our proposed schemes.ROBUSTNESS AND TRACTABILITY FOR NONCONVEX M-ESTIMATORS
AbstractZhang, R., Mei, Y., Shi, J., & Xu, H. (n.d.).Publication year
2022Journal title
Statistica SinicaVolume
32Issue
3Page(s)
1295-1316AbstractWe investigate two important properties of M-estimators, namely, robustness and tractability, in a linear regression setting, when the observations are contaminated by some arbitrary outliers. Specifically, robustness is the statistical property that the estimator should always be close to the true underlying parameters, regardless of the distribution of the outliers, and tractability refers to the computational property that the estimator can be computed efficiently, even if the objective function of the M-estimator is nonconvex. In this article, by examining the empirical risk, we show that under some sufficient conditions, many M-estimators enjoy nice robustness and tractability properties simultaneously when the percentage of outliers is small. We extend our analysis to the high-dimensional setting, where the number of parameters is greater than the number of samples, p ≫ n, and prove that when the proportion of outliers is small, the penalized M-estimators with the L1 penalty enjoy robustness and tractability simultaneously. Our research provides an analytic approach to determine the effects of outliers and tuning parameters on the robustness and tractability of some families of M-estimators. Simulations and case studies are presented to illustrate the usefulness of our theoretical results for M-estimators under Welsch’s exponential squared loss and Tukey’s bisquare loss.The Directional Bias Helps Stochastic Gradient Descent to Generalize in Kernel Regression Models
AbstractLuo, Y., Huo, X., & Mei, Y. (n.d.).Publication year
2022Page(s)
678-683AbstractWe study the Stochastic Gradient Descent (SGD) algorithm in nonparametric statistics: kernel regression in particular. The directional bias property of SGD, which is known in the linear regression setting, is generalized to the kernel regression. More specifically, we prove that SGD with moderate and annealing step-size converges along the direction of the eigenvector that corresponds to the largest eigenvalue of the Gram matrix. In addition, the Gradient Descent (GD) with a moderate or small step-size converges along the direction that corresponds to the smallest eigenvalue. These facts are referred to as the directional bias properties; they may interpret how an SGD-computed estimator has a potentially smaller generalization error than a GD-computed estimator. The application of our theory is demonstrated by simulation studies and a case study that is based on the FashionMNIST dataset.Treatment Effect Modeling for FTIR Signals Subject to Multiple Sources of Uncertainties
AbstractTian, H., Wang, A., Chen, J., Jiang, X., Shi, J., Zhang, C., Mei, Y., & Wang, B. (n.d.).Publication year
2022Journal title
IEEE Transactions on Automation Science and EngineeringVolume
19Issue
2Page(s)
895-906AbstractFourier-transform infrared spectroscopy (FTIR) is a widely adopted technique for characterizing the chemical composition in many physical and chemical analyses. However, FTIR spectra are subject to multiple sources of uncertainty, and thus the analysis of them relies on domain experts and can only lead to qualitative conclusions. This study aims to analyze the effect of a certain treatment on FTIR spectra subject to two commonly observed uncertainties, the offset shift and the multiplicative error. Due to these uncertainties, the pre-exposure FTIR spectra are modeled according to the physical understanding of the uncertainty - observed spectra can be viewed as translating and stretchering an underlying template signal, and the post-exposure FTIR spectra are modeled as the translated and stretchered template signal plus an extra functional treatment effect. To provide engineering interpretation, the treatment effect is modeled as the product of the pattern of modification and its corresponding magnitude. A two-step parameter estimation algorithm is developed to estimate the underlying template signal, the pattern of modification, and the magnitude of modification at various treatment strengths. The effectiveness of the proposed method is validated in a simulation study. Furtherly, in a real case study, the proposed method is used to investigate the effect of plasma exposure on the FTIR spectra. As a result, the proposed method effectively identifies the pattern of modification under uncertainties in the manufacturing environment, which matches the knowledge of the affected chemical components by the plasma treatment. And the recovered magnitude of modification provides guidance in selecting the control parameter of the plasma treatment. Note to Practitioners - FTIR spectrometer is often used to characterize the surface chemical composition of a material. Due to the large uncertainties associated with the nature of spectrometer and the measurement environment, the FTIR signals are usually examined visually by experienced engineers and technicians in industrial applications, which can be both time-consuming and inaccurate. To understand the effect of plasma exposure on the surface property of carbon fiber reinforced polymer (CFRP) material, the elimination of uncertainties associated with FTIR signals is investigated, and a systematic method is proposed to quantify the effect of surface treatments on FTIR signals. A two-step analytic procedure is proposed, which provides information on how the plasma exposure distorts the FTIR signals, and how the plasma distance relates to the magnitude of the distortion. The methodology in this article can be used to analyze the treatment effect on a variety of spectroscopic measurements that are subject to uncertainties such as offset and scaling errors, which expands the applications of in situ handheld spectrometer metrology in manufacturing industries.A boosting inspired personalized threshold method for sepsis screening
AbstractFeng, C., Feng, C., Griffin, P., Kethireddy, S., & Mei, Y. (n.d.).Publication year
2021Journal title
Journal of Applied StatisticsVolume
48Issue
1Page(s)
154-175AbstractSepsis is one of the biggest risks to patient safety, with a natural mortality rate between 25% and 50%. It is difficult to diagnose, and no validated standard for diagnosis currently exists. A commonly used scoring criteria is the quick sequential organ failure assessment (qSOFA). It demonstrates very low specificity in ICU populations, however. We develop a method to personalize thresholds in qSOFA that incorporates easily to measure patient baseline characteristics. We compare the personalized threshold method to qSOFA, five previously published methods that obtain an optimal constant threshold for a single biomarker, and to the machine learning algorithms based on logistic regression and AdaBoosting using patient data in the MIMIC-III database. The personalized threshold method achieves higher accuracy than qSOFA and the five published methods and has comparable performance to machine learning methods. Personalized thresholds, however, are much easier to adopt in real-life monitoring than machine learning methods as they are computed once for a patient and used in the same way as qSOFA, whereas the machine learning methods are hard to implement and interpret.Aneurysmal Subarachnoid Hemorrhage : Trends, Outcomes, and Predictions from a 15-Year Perspective of a Single Neurocritical Care Unit
AbstractSamuels, O. B., Sadan, O., Feng, C., Feng, C., Martin, K., Medani, K., Mei, Y., & Barrow, D. L. (n.d.).Publication year
2021Journal title
NeurosurgeryVolume
88Issue
3Page(s)
574-583AbstractBACKGROUND: Aneurysmal subarachnoid hemorrhage (aSAH) is associated with disproportionally high mortality and long-term neurological sequelae. Management of patients with aSAH has changed markedly over the years, leading to improvements in outcome. OBJECTIVE: To describe trends in aSAH care and outcome in a high-volume single center 15-yr cohort. METHODS: All new admissions diagnosed with subarachnoid hemorrhage (SAH) to our tertiary neuro-intensive care unit between 2002 and 2016 were reviewed. Trend analysis was performed to assess temporal changes and a step-wise regression analysis was done to identify factors associated with outcomes. RESULTS: Out of 3970 admissions of patients with SAH, 2475 patients proved to have a ruptured intracranial aneurysm. Over the years of the study, patient acuity increased by Hunt & Hess (H&H) grade and related complications. Endovascular therapies became more prevalent over the years, and were correlated with better outcome. Functional outcome overall improved, yet the main effect was noted in the low- and intermediate-grade patients. Several parameters were associated with poor functional outcome, including long-term mechanical ventilation (odds ratio 11.99, CI 95% [7.15-20.63]), acute kidney injury (3.55 [1.64-8.24]), pneumonia (2.89 [1.89-4.42]), hydrocephalus (1.80 [1.24-2.63]) diabetes mellitus (1.71 [1.04-2.84]), seizures (1.69 [1.07-2.70], H&H (1.67 [1.45-1.94]), and age (1.06 [1.05-1.07]), while endovascular approach to treat the aneurysm, compared with clip-ligation, had a positive effect (0.35 [0.25-0.48]). CONCLUSION: This large, single referral center, retrospective analysis reveals important trends in the treatment of aSAH. It also demonstrates that despite improvement in functional outcome over the years, systemic complications remain a significant risk factor for poor prognosis. The historic H&H determination of outcome is less valid with today's improved care.Correlation-based dynamic sampling for online high dimensional process monitoring
AbstractNabhan, M., Mei, Y., & Shi, J. (n.d.).Publication year
2021Journal title
Journal of Quality TechnologyVolume
53Issue
3Page(s)
289-308AbstractEffective process monitoring of high-dimensional data streams with embedded spatial structures has been an arising challenge for environments with limited resources. Utilizing the spatial structure is key to improve monitoring performance. This article proposes a correlation-based dynamic sampling technique for change detection. Our method borrows the idea of Upper Confidence Bound algorithm and uses the correlation structure not only to calculate a global statistic, but also to infer unobserved sensors from partial observations. Simulation studies and two case studies on solar flare detection and carbon nanotubes (CNTs) buckypaper process monitoring are used to validate the effectiveness of our method.Creation of a Pediatric Choledocholithiasis Prediction Model
AbstractCohen, R. Z., Tian, H., Sauer, C. G., Willingham, F. F., Santore, M. T., Mei, Y., & Freeman, A. J. (n.d.).Publication year
2021Journal title
Journal of Pediatric Gastroenterology and NutritionVolume
73Issue
5Page(s)
636-641AbstractBackground:Definitive non-invasive detection of pediatric choledocholithiasis could allow more efficient identification of those patients who are most likely to benefit from therapeutic endoscopic retrograde cholangiopancreatography (ERCP) for stone extraction.Objective:To craft a pediatric choledocholithiasis prediction model using a combination of commonly available serum laboratory values and ultrasound results.Methods:A retrospective review of laboratory and imaging results from 316 pediatric patients who underwent intraoperative cholangiogram or ERCP due to suspicion of choledocholithiasis were collected and compared to presence of common bile duct stones on cholangiography. Multivariate logistic regression with supervised machine learning was used to create a predictive scoring model. Monte-Carlo cross-validation was used to validate the scoring model and a score threshold that would provide at least 90% specificity for choledocholithiasis was determined in an effort to minimize non-therapeutic ERCP.Results:Alanine aminotransferase (ALT), total bilirubin, alkaline phosphatase, and common bile duct diameter via ultrasound were found to be the key clinical variables to determine the likelihood of choledocholithiasis. The dictated specificity threshold of 90.3% yielded a sensitivity of 40.8% and overall accuracy of 71.5% in detecting choledocholithiasis. Positive predictive value was 71.4% and negative predictive value was 72.1%.Conclusion:Our novel pediatric choledocholithiasis predictive model is a highly specific tool to suggest ERCP in the setting of likely choledocholithiasis.Editorial : Mathematical Fundamentals of Machine Learning
AbstractGlickenstein, D., Hamm, K., Huo, X., Mei, Y., & Stoll, M. (n.d.).Publication year
2021Journal title
Frontiers in Applied Mathematics and StatisticsVolume
7Abstract~Multi-Stream Quickest Detection with Unknown Post-Change Parameters under Sampling Control
AbstractAbstractThe multi-stream quickest detection problem with unknown post-change parameters is studied under the sampling control constraint, where there are M local processes in a system but one is only able to take observations from one of these M local processes at each time instant. The objective is to raise a correct alarm as quickly as possible once the change occurs subject to both false alarm and sampling control constraints. We propose an efficient myopic-sampling-based quickest detection algorithm under sampling control constraint, and show it is asymptotically optimal in the sense of minimizing the detection delay under our context when the number M of processes is fixed. Simulation studies are conducted to validate our theoretical results.