A5011089052- Control Systems and Identification
- Fuzzy Systems and Optimization
- Multi-Criteria Decision Making
- Advanced biosensing and bioanalysis techniques
- Financial Risk and Volatility Modeling
- Optimization and Search Problems
- Advanced Proteomics Techniques and Applications
- Numerical methods in inverse problems
- Advanced Statistical Methods and Models
- Advanced Bandit Algorithms Research
- Risk and Portfolio Optimization
- Probabilistic and Robust Engineering Design
- Engineering Applied Research
- Machine Learning and Algorithms
- COVID-19 epidemiological studies
- Fault Detection and Control Systems
- Refrigeration and Air Conditioning Technologies
- Building Energy and Comfort Optimization
- Statistical Methods and Bayesian Inference
- Income, Poverty, and Inequality
- Statistical Methods and Inference
- Advanced Nanomaterials in Catalysis
- demographic modeling and climate adaptation
Columbia University
2022-2024
Beijing University of Technology
2022
Guangdong University of Petrochemical Technology
2022
We construct a peptide-conjugated metal cluster as an enzyme-like catalytic bioprobe to enhance quantitative analysis of membrane protein biomarker and detect epithelial-to-mesenchymal transition tumor cells. This with atomically precise formula, termed clusterzyme, possesses selective recognition intrinsic activity. These favorable features facilitate sensitive the in situ through on-cell signal amplification. clusterzyme-based analytical method exhibits excellent compatibility traditional...
In this article, we propose an optimal predictor of a random variable that has either infinite mean or variance. The method consists transforming the such transformed finite and proposed is generalized arithmetic which similar to notion certainty price in utility theory. Typically, transformation parametric family bijections, case parameter might be chosen minimize prediction error coordinates. statistical properties estimator are studied, confidence intervals provided. performance procedure...
Despite a large and significant body of recent work focused on estimating the out-of-sample risk regularized models in high dimensional regime, theoretical understanding this problem for non-differentiable penalties such as generalized LASSO nuclear norm is missing. In paper we resolve challenge. We study proportional regime where both sample size n number features p are large, n/p signal-to-noise ratio (per observation) remain finite. provide finite upper bounds expected squared error...
In this paper we propose an optimal predictor of a random variable that has either infinite mean or variance. The method consists transforming the such transformed finite and proposed is generalized arithmetic which similar to notion certainty price in utility theory. Typically, transformation parametric family bijections, case parameter might be chosen minimize prediction error coordinates. statistical properties estimator are studied, confidence intervals provided. performance procedure...
The out-of-sample error (OO) is the main quantity of interest in risk estimation and model selection. Leave-one-out cross validation (LO) offers a (nearly) distribution-free yet computationally demanding approach to estimate OO. Recent theoretical work showed that approximate leave-one-out (ALO) efficient statistically reliable LO (and OO) for generalized linear models with differentiable regularizers. For problems involving non-differentiable regularizers, despite significant empirical...
Online learning to rank is a core problem in machine learning. In Lattimore et al. (2018), novel online algorithm was proposed based on topological sorting. the paper they provided set of self-normalized inequalities (a) as criterion iterations and (b) provide an upper bound for cumulative regret, which measure performance. this work, we utilized method mixtures asymptotic expansions certain implicit function tighter, iterated-log-like boundary inequalities, consequence improve both itself...
Abstract Hitting time, or first passage is the time that a process reaches (from above below) pre-specified area. Estimating expected hitting an important practical problem in many areas of science. In this paper, we discuss novel and approach to estimating [b, ∞), for boundary level b, as function inverse existing projections. An example are predictions from ensemble models. Convergence properties estimator established. Applications epidemiology climatology presented. The performance...
This study presents a movable electrical heating radiator with the function of user's temperature tracing and constant control to enhance thermal comfortable users. The system was modified by previous studies in our laboratory. unit engaged Holtek HT66F2390 which is newly designed microcontroller China. sensing camera detected transferred data move follow user up. results display reliable performance operations. better distance between traditional heat 150 250 cm. However, might too long...