Yichao Wu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Yichao Wu has a total of 98 co-authors. Here are the top co-authors by number of shared publications:
- Yufeng Liu
Department of Statistics, University of Michigan
7 shared publications Chaowen Zheng
Department of Statistics North Carolina State University Raleigh North Carolina USA
3 shared publications- Hao Helen Zhang
Statistics and Data Science GIDP University of Arizona 85721 Tucson AZ USA
3 shared publications - Hua Zhou
Department of Biostatistics, University of California, Los Angeles, Los Angeles, USA
3 shared publications - Jianqing Fan
Department of Operations Research and Financial Engineering, Princeton University, Princeton, New Jersey 08544
3 shared publications - Jingfang Huang
The University of North Carolina at Chapel Hill
3 shared publications
Latest Publications
- Open Access
Quadratic Formula-based Nonlinear Approximation
This paper presents a quadratic formula-based nonlinear representation for a given single-variable function f(x), $-1 \leq x \leq 1$. First, we construct the explicit polynomial coefficient functions a(x), b(x), and c(x) using a least-squares approach. Then, f is reconstructed by solving the degree-2 polynomial equation a(x) f^2 - b(x) f - c(x)=0 for any $x \in [-1,1]$, where an index function is used to select the correct sign in the quadratic formula. The quadratic formula-based nonlinear approximation (degree-2 in f) outperforms classical orthogonal polynomial-based least-squares approximation (degree-0 in f) and rational approximation (degree-1 in f) for functions with sharp transitions or discontinuities. As a potential application, we apply the degree-2 representation to data denoising. Instead of relying on more complex "edge-preserving" metric-based optimization techniques, the smooth coefficient functions a(x), b(x), and c(x) enable effective least-squares-based denoising on the low-dimensional manifold described by the algebraic variety a(x) f^2 - b(x) f - c(x)=0. Denoising the index function, which determines the appropriate root to select, can be achieved using classical statistical or modern classification/clustering techniques. Numerical results and data denoising examples are provided to demonstrate the effectiveness of the degree-2 nonlinear approximation technique. The new nonlinear, quadratic formula-based representation also raises theoretical and numerical questions, including strategies for identifying numerically stable representations, developing optimal algorithms to construct the polynomial coefficient functions a(x), b(x), and c(x), and achieving economical representation and denoising of the index function.Open AccessQuadratic Formula-based Nonlinear Approximation
This paper presents a quadratic formula-based nonlinear representation for a given single-variable function f(x), $-1 \leq x \leq 1$. First, we construct the explicit polynomial coefficient functions a(x), b(x), and c(x) using a least-squares approach. Then, f is reconstructed by solving the degree-2 polynomial equation a(x) f^2 - b(x) f - c(x)=0 for any $x \in [-1,1]$, where an index function is used to select the correct sign in the quadratic formula. The quadratic formula-based nonlinear approximation (degree-2 in f) outperforms classical orthogonal polynomial-based least-squares approximation (degree-0 in f) and rational approximation (degree-1 in f) for functions with sharp transitions or discontinuities. As a potential application, we apply the degree-2 representation to data denoising. Instead of relying on more complex "edge-preserving" metric-based optimization techniques, the smooth coefficient functions a(x), b(x), and c(x) enable effective least-squares-based denoising on the low-dimensional manifold described by the algebraic variety a(x) f^2 - b(x) f - c(x)=0. Denoising the index function, which determines the appropriate root to select, can be achieved using classical statistical or modern classification/clustering techniques. Numerical results and data denoising examples are provided to demonstrate the effectiveness of the degree-2 nonlinear approximation technique. The new nonlinear, quadratic formula-based representation also raises theoretical and numerical questions, including strategies for identifying numerically stable representations, developing optimal algorithms to construct the polynomial coefficient functions a(x), b(x), and c(x), and achieving economical representation and denoising of the index function. - Open Access
Basis Choices for Frequency Domain Statistical Independence Tests and Algorithms for Algebraic Relation Extraction
In this paper, we explore how different selections of basis functions impact the efficacy of frequency domain techniques in statistical independence tests, and study different algorithms for extracting low-dimensional algebraic relations from dependent data. We examine a range of complete orthonormal bases functions including the Legendre polynomials, Fourier series, Walsh functions, and standard and nonstandard Haar wavelet bases. We utilize fast transformation algorithms to efficiently transform physical domain data to frequency domain coefficients. The main focuses of this paper are the effectiveness of different basis selections in detecting data dependency using frequency domain data, e.g., whether varying basis choices significantly influence statistical power loss for small data with large noise; and on the stability of different optimization formulations for finding proper algebraic relations when data are dependent. We present numerical results to demonstrate the effectiveness of frequency domain-based statistical analysis methods and provide guidance for selecting the proper basis and algorithm to detect a particular type of relations.
Open AccessBasis Choices for Frequency Domain Statistical Independence Tests and Algorithms for Algebraic Relation Extraction
In this paper, we explore how different selections of basis functions impact the efficacy of frequency domain techniques in statistical independence tests, and study different algorithms for extracting low-dimensional algebraic relations from dependent data. We examine a range of complete orthonormal bases functions including the Legendre polynomials, Fourier series, Walsh functions, and standard and nonstandard Haar wavelet bases. We utilize fast transformation algorithms to efficiently transform physical domain data to frequency domain coefficients. The main focuses of this paper are the effectiveness of different basis selections in detecting data dependency using frequency domain data, e.g., whether varying basis choices significantly influence statistical power loss for small data with large noise; and on the stability of different optimization formulations for finding proper algebraic relations when data are dependent. We present numerical results to demonstrate the effectiveness of frequency domain-based statistical analysis methods and provide guidance for selecting the proper basis and algorithm to detect a particular type of relations.
- Open Access
- 25 December 2023
- Published by Wiley in Alzheimer's & Dementia
- Vol. 19 (S12)
Signatures of abnormal deviations in excitation‐inhibition balance for amyloid positive individuals
Background: Dysfunctional excitation‐inhibition balance (EIB) is hypothesized to precede cognitive impairment in Alzheimer’s disease (AD); however, overlapping neuropathology (like amyloid plaques) can develop in cognitively normal (CN) and impaired (CI) individuals. We investigate EIB trajectories for individuals with significant amyloid (Aβ+).Method: Multimodal imaging was used to construct hybrid resting‐state structural connectomes (Fortel, Igor, et al. Network Neuroscience 6.2 (2022): 420‐444.), computing whole‐brain EIB based on 105 ROIs. This cohort (N = 185, 94 Female) aged 55‐90 (μ FEMALES = 74.9, and μ MALES 73) was limited to Aβ+ individuals with APOE genotyping (ε3/ε3, ε3/ε4, ε4/ε4); grouped further by clinical dementia rating (CDR), with CDR > 0 indicating CI. Imaging, Aβ+ classifications, CDR, and APOE were obtained from OASIS‐3: Longitudinal Multimodal Neuroimaging: Principal Investigators: T. Benzinger, D. Marcus, J. Morris; NIH P30 AG066444, P50 AG00561, P30 NS09857781, P01 AG026276, P01 AG003991, R01 AG043434, UL1 TR000448, R01 EB009352. AV‐45 doses were provided by Avid Radiopharmaceuticals, a wholly owned subsidiary of Eli Lilly.Result: Results presented in Figure 1 suggest whole‐brain EIB in both CN sexes with Aβ+ decreases with increasing age (APOE‐ε4 agnostic). However, sex differences are highlighted with CI; regression analyses produce a marginally significant 4‐way interaction of age, sex, cognitive status, and APOE (P = 0.085, β = ‐0.013). Partial correlation of EIB with age, controlling for sex, APOE and cognitive status is also significant (P = 0.016, r = ‐0.18). Results indicate sex‐mediated EIB differences in individuals with CI, dominated by APOE‐ε4 in women rather than amyloid.Conclusion: We’ve demonstrated EIB deviations in cognitively impaired men and women that may be associated with AD neuropathology. Our findings are consistent with known AD sex differences and suggest novel multimodal connectomics may capture subtle disruptions in brain dynamics that traditional methods may not. Figure. 1. Whole‐brain excitation‐Inhibition balance (EIB) trajectories for amyloid positive (Aβ+) men and women. The cognitively normal (CN) group (rows A‐B) suggests no sex differences in EIB trajectories with age, while the cognitively impaired (CI) group (rows C‐D) begins to reveal distinct patterns of dysfunction (more significantly in women with APOE‐ε4). Note a value of 0.5 indicates perfect balance.
- 25 December 2023
- Published by Wiley in Alzheimer's & Dementia
- Vol. 19 (S12)
Open AccessSignatures of abnormal deviations in excitation‐inhibition balance for amyloid positive individuals
Background: Dysfunctional excitation‐inhibition balance (EIB) is hypothesized to precede cognitive impairment in Alzheimer’s disease (AD); however, overlapping neuropathology (like amyloid plaques) can develop in cognitively normal (CN) and impaired (CI) individuals. We investigate EIB trajectories for individuals with significant amyloid (Aβ+).Method: Multimodal imaging was used to construct hybrid resting‐state structural connectomes (Fortel, Igor, et al. Network Neuroscience 6.2 (2022): 420‐444.), computing whole‐brain EIB based on 105 ROIs. This cohort (N = 185, 94 Female) aged 55‐90 (μ FEMALES = 74.9, and μ MALES 73) was limited to Aβ+ individuals with APOE genotyping (ε3/ε3, ε3/ε4, ε4/ε4); grouped further by clinical dementia rating (CDR), with CDR > 0 indicating CI. Imaging, Aβ+ classifications, CDR, and APOE were obtained from OASIS‐3: Longitudinal Multimodal Neuroimaging: Principal Investigators: T. Benzinger, D. Marcus, J. Morris; NIH P30 AG066444, P50 AG00561, P30 NS09857781, P01 AG026276, P01 AG003991, R01 AG043434, UL1 TR000448, R01 EB009352. AV‐45 doses were provided by Avid Radiopharmaceuticals, a wholly owned subsidiary of Eli Lilly.Result: Results presented in Figure 1 suggest whole‐brain EIB in both CN sexes with Aβ+ decreases with increasing age (APOE‐ε4 agnostic). However, sex differences are highlighted with CI; regression analyses produce a marginally significant 4‐way interaction of age, sex, cognitive status, and APOE (P = 0.085, β = ‐0.013). Partial correlation of EIB with age, controlling for sex, APOE and cognitive status is also significant (P = 0.016, r = ‐0.18). Results indicate sex‐mediated EIB differences in individuals with CI, dominated by APOE‐ε4 in women rather than amyloid.Conclusion: We’ve demonstrated EIB deviations in cognitively impaired men and women that may be associated with AD neuropathology. Our findings are consistent with known AD sex differences and suggest novel multimodal connectomics may capture subtle disruptions in brain dynamics that traditional methods may not. Figure. 1. Whole‐brain excitation‐Inhibition balance (EIB) trajectories for amyloid positive (Aβ+) men and women. The cognitively normal (CN) group (rows A‐B) suggests no sex differences in EIB trajectories with age, while the cognitively impaired (CI) group (rows C‐D) begins to reveal distinct patterns of dysfunction (more significantly in women with APOE‐ε4). Note a value of 0.5 indicates perfect balance.
- Open Access
- 18 October 2022
- Published by Wiley in Canadian Journal of Statistics
- Vol. 51 (4) , 959-974
Automatic structure recovery for generalized additive models
- 18 October 2022
- Published by Wiley in Canadian Journal of Statistics
- Vol. 51 (4) , 959-974
Open AccessAutomatic structure recovery for generalized additive models
- Open Access
- 16 April 2022
- Published by Wiley in Mathematical Problems in Engineering
- Vol. 2022, 1-9
Rural Tourism Public Service Performance Evaluation Based on Analytic Hierarchy Process
In recent years, more and more attention has been paid to the construction of rural tourism, and tourists’ demand for rural tourism public services has become stronger and stronger. Through the objective and fair evaluation of rural tourism public service performance, on the one hand, it is helpful to improve the overall satisfaction of tourists to rural tourism. On the other hand, it can also ensure the efficiency of public service financial investment and expenditure. Firstly, on the basis of theoretical summary and hierarchical analysis, this paper constructs the performance evaluation index system of rural tourism public service from five aspects of infrastructure services, public transport services, public information services, security services, and public environmental services. Then taking China as an example, the performance of rural tourism public service is evaluated by analytic hierarchy process. The results show that the construction of hard environment such as public transport services, security services, and infrastructure services is a key factor in the construction of rural tourism public services. Finally, according to the evaluation results obtained in this paper, some suggestions are put forward for the development of rural tourism and the improvement of public service quality in the future.
- 16 April 2022
- Published by Wiley in Mathematical Problems in Engineering
- Vol. 2022, 1-9
Open AccessRural Tourism Public Service Performance Evaluation Based on Analytic Hierarchy Process
In recent years, more and more attention has been paid to the construction of rural tourism, and tourists’ demand for rural tourism public services has become stronger and stronger. Through the objective and fair evaluation of rural tourism public service performance, on the one hand, it is helpful to improve the overall satisfaction of tourists to rural tourism. On the other hand, it can also ensure the efficiency of public service financial investment and expenditure. Firstly, on the basis of theoretical summary and hierarchical analysis, this paper constructs the performance evaluation index system of rural tourism public service from five aspects of infrastructure services, public transport services, public information services, security services, and public environmental services. Then taking China as an example, the performance of rural tourism public service is evaluated by analytic hierarchy process. The results show that the construction of hard environment such as public transport services, security services, and infrastructure services is a key factor in the construction of rural tourism public services. Finally, according to the evaluation results obtained in this paper, some suggestions are put forward for the development of rural tourism and the improvement of public service quality in the future.
- 1 January 2022
- Published by Statistica Sinica (Institute of Statistical Science) in Statistica Sinica
- Vol. 32 (3) , 1563-1582
Nonparametric Interaction Selection
- 1 January 2022
- Published by Statistica Sinica (Institute of Statistical Science) in Statistica Sinica
- Vol. 32 (3) , 1563-1582
Nonparametric Interaction Selection
- Open Access
Knowledge Distillation for Object Detection via Rank Mimicking and Prediction-guided Feature Imitation
Knowledge Distillation (KD) is a widely-used technology to inherit information from cumbersome teacher models to compact student models, consequently realizing model compression and acceleration. Compared with image classification, object detection is a more complex task, and designing specific KD methods for object detection is non-trivial. In this work, we elaborately study the behaviour difference between the teacher and student detection models, and obtain two intriguing observations: First, the teacher and student rank their detected candidate boxes quite differently, which results in their precision discrepancy. Second, there is a considerable gap between the feature response differences and prediction differences between teacher and student, indicating that equally imitating all the feature maps of the teacher is the sub-optimal choice for improving the student's accuracy. Based on the two observations, we propose Rank Mimicking (RM) and Prediction-guided Feature Imitation (PFI) for distilling one-stage detectors, respectively. RM takes the rank of candidate boxes from teachers as a new form of knowledge to distill, which consistently outperforms the traditional soft label distillation. PFI attempts to correlate feature differences with prediction differences, making feature imitation directly help to improve the student's accuracy. On MS COCO and PASCAL VOC benchmarks, extensive experiments are conducted on various detectors with different backbones to validate the effectiveness of our method. Specifically, RetinaNet with ResNet50 achieves 40.4% mAP in MS COCO, which is 3.5% higher than its baseline, and also outperforms previous KD methods.
Open AccessKnowledge Distillation for Object Detection via Rank Mimicking and Prediction-guided Feature Imitation
Knowledge Distillation (KD) is a widely-used technology to inherit information from cumbersome teacher models to compact student models, consequently realizing model compression and acceleration. Compared with image classification, object detection is a more complex task, and designing specific KD methods for object detection is non-trivial. In this work, we elaborately study the behaviour difference between the teacher and student detection models, and obtain two intriguing observations: First, the teacher and student rank their detected candidate boxes quite differently, which results in their precision discrepancy. Second, there is a considerable gap between the feature response differences and prediction differences between teacher and student, indicating that equally imitating all the feature maps of the teacher is the sub-optimal choice for improving the student's accuracy. Based on the two observations, we propose Rank Mimicking (RM) and Prediction-guided Feature Imitation (PFI) for distilling one-stage detectors, respectively. RM takes the rank of candidate boxes from teachers as a new form of knowledge to distill, which consistently outperforms the traditional soft label distillation. PFI attempts to correlate feature differences with prediction differences, making feature imitation directly help to improve the student's accuracy. On MS COCO and PASCAL VOC benchmarks, extensive experiments are conducted on various detectors with different backbones to validate the effectiveness of our method. Specifically, RetinaNet with ResNet50 achieves 40.4% mAP in MS COCO, which is 3.5% higher than its baseline, and also outperforms previous KD methods.
- 27 September 2021
- Published by Taylor & Francis in Journal of the American Statistical Association
- Vol. 118 (542) , 1023-1037
Variable Selection for Global Fréchet Regression
Global Fréchet regression is an extension of linear regression to cover more general types of responses, such as distributions, networks, and manifolds, which are becoming more prevalent. In such models, predictors are Euclidean while responses are metric space valued. Predictor selection is of major relevance for regression modeling in the presence of multiple predictors but has not yet been addressed for Fréchet regression. Due to the metric space-valued nature of the responses, Fréchet regression models do not feature model parameters, and this lack of parameters makes it a major challenge to extend existing variable selection methods for linear regression to global Fréchet regression. In this work, we address this challenge and propose a novel variable selection method that overcomes it and has good practical performance. We provide theoretical support and demonstrate that the proposed variable selection method achieves selection consistency. We also explore the finite sample performance of the proposed method with numerical examples and data illustrations.
- 27 September 2021
- Published by Taylor & Francis in Journal of the American Statistical Association
- Vol. 118 (542) , 1023-1037
Variable Selection for Global Fréchet Regression
Global Fréchet regression is an extension of linear regression to cover more general types of responses, such as distributions, networks, and manifolds, which are becoming more prevalent. In such models, predictors are Euclidean while responses are metric space valued. Predictor selection is of major relevance for regression modeling in the presence of multiple predictors but has not yet been addressed for Fréchet regression. Due to the metric space-valued nature of the responses, Fréchet regression models do not feature model parameters, and this lack of parameters makes it a major challenge to extend existing variable selection methods for linear regression to global Fréchet regression. In this work, we address this challenge and propose a novel variable selection method that overcomes it and has good practical performance. We provide theoretical support and demonstrate that the proposed variable selection method achieves selection consistency. We also explore the finite sample performance of the proposed method with numerical examples and data illustrations.
- 5 August 2021
- Published by Wiley in Canadian Journal of Statistics
- Vol. 50 (2) , 612-637
A modified expectation‐maximization algorithm for latent Gaussian graphical model
This paper considers the latent Gaussian graphical model, which extends the Gaussian graphical model to handle discrete data as well as mixed data with both continuous and discrete variables by assuming that discrete variables are generated by discretizing latent Gaussian variables. We propose a modified expectation‐maximization (EM) algorithm to estimate parameters in the latent Gaussian model for binary data. We also extend the proposed modified EM algorithm to the latent Gaussian model for mixed data. The conditional dependence structure can be consequently constructed by exploring the sparsity pattern of the precision matrix of the latent variables. We illustrate the performance of our proposed estimator through comprehensive numerical studies and an application to voting data of the United Nations General Assembly.
- 5 August 2021
- Published by Wiley in Canadian Journal of Statistics
- Vol. 50 (2) , 612-637
A modified expectation‐maximization algorithm for latent Gaussian graphical model
This paper considers the latent Gaussian graphical model, which extends the Gaussian graphical model to handle discrete data as well as mixed data with both continuous and discrete variables by assuming that discrete variables are generated by discretizing latent Gaussian variables. We propose a modified expectation‐maximization (EM) algorithm to estimate parameters in the latent Gaussian model for binary data. We also extend the proposed modified EM algorithm to the latent Gaussian model for mixed data. The conditional dependence structure can be consequently constructed by exploring the sparsity pattern of the precision matrix of the latent variables. We illustrate the performance of our proposed estimator through comprehensive numerical studies and an application to voting data of the United Nations General Assembly.