A5027388649- Advanced Queuing Theory Analysis
- Advanced Bandit Algorithms Research
- Simulation Techniques and Applications
- Statistical Methods and Inference
- Risk and Portfolio Optimization
- Supramolecular Chemistry and Complexes
- Auction Theory and Applications
- Advanced Statistical Methods and Models
- Hydrocarbon exploration and reservoir analysis
- Machine Learning and Algorithms
- Machine Learning in Healthcare
- Consumer Market Behavior and Pricing
- Stochastic processes and financial applications
- Reinforcement Learning in Robotics
- Coal Properties and Utilization
- Crystallization and Solubility Studies
- Gaussian Processes and Bayesian Inference
- Molecular Sensors and Ion Detection
- Fault Detection and Control Systems
- Random Matrices and Applications
- X-ray Diffraction in Crystallography
- Enhanced Oil Recovery Techniques
- Insurance, Mortality, Demography, Risk Management
- Decision-Making and Behavioral Economics
- Explainable Artificial Intelligence (XAI)
Anhui University of Technology
2023-2025
Henan Polytechnic University
2022-2025
University of Hong Kong
2024
Hong Kong University of Science and Technology
2024
Stanford University
2019-2023
University of Chicago
2023
Zhejiang Shuren University
2022
Peking University
2016
Heritage Christian University
2010
Abstract Although the synthesis of mechanically interlocked molecules has been extensively researched, selectively constructing homogeneous linear [4]catenanes remains a formidable challenge. Here, we constructed metalla[4]catenane in one‐step process through coordination‐driven self‐assembly bidentate benzothiadiazole derivative ligand and binuclear half‐sandwich rhodium precursor. The formation metalla[4]catenanes was facilitated by cooperative interactions between strong sandwich‐type π‐π...
Randomized experiments, or A/B testing, are the gold standard for evaluating interventions but underutilized in area of inventory management. This study addresses this gap by analyzing testing strategies multi-item, multi-period systems with lost sales and capacity constraints. We examine switchback item-level randomization, pairwise staggered rollouts, their biases theoretically comparing them through numerical experiments. Our findings provide actionable guidance selecting experimental...
Transfer learning is a popular strategy to leverage external knowledge and improve statistical efficiency, particularly with limited target sample. We propose novel knowledge-guided Wasserstein Distributionally Robust Optimization (KG-WDRO) framework that adaptively incorporates multiple sources of overcome the conservativeness vanilla WDRO, which often results in overly pessimistic shrinkage toward zero. Our method constructs smaller ambiguity sets by controlling transportation along...
In this study, the normalized fractal dimension (DN) model of full-scale pore size was established based on classical scaling relationship porous materials. The methodology described in detail, and rationality examined by between volume specific surface area (SSA). results indicate that is a continuous function full scale, which can more comprehensively symbolize characteristic distribution scale. addition, quantitatively characterize absolute compared with traditional segmented relatively...
Despite substantial advancements in the synthesis of mechanically interlocked molecules (MIMs), achieving efficient construction higher‐order links remains a formidable challenge. Herein, we report highly one‐step directed series unprecedented molecular closed four‐link chains (84 1 metalla‐links), achieved through synergistic assembly coordination‐driven and aromatic stacking interactions involving binuclear rhodium/iridium precursors bidentate benzothiadiazole derivative ligands....
Despite substantial advancements in the synthesis of mechanically interlocked molecules (MIMs), achieving efficient construction higher‐order links remains a formidable challenge. Herein, we report highly one‐step directed series unprecedented molecular closed four‐link chains (84 1 metalla‐links), achieved through synergistic assembly coordination‐driven and aromatic stacking interactions involving binuclear rhodium/iridium precursors bidentate benzothiadiazole derivative ligands....
In this study, we conduct laboratory experiments on coal with liquid CO2 phase transition fracturing (L-CO2-PTF) treatment under the pressures of 120 and 185 MPa. The variations structure fractal characteristics for mesopores (2–50 nm) micropores (<2 are studied by employing low-temperature N2/CO2 adsorption measurements theory. results indicate that effects pore enlarging dimension reducing L-CO2-PTF visible mesopores, while these underperformed micropores. connectivity is improved due to...
In this work, we have given an analogical method for estimating the fractal dimension three-dimensional fracture tortuosity (3D-FT). The comparison and error analysis of rigorous methods on 3D-FT were carried out in work. [Formula: see text] from proposed is function 3D average ([Formula: length text]. with high accuracy indicates good consistency embodiment physical meaning relatively convenient calculating premise ensuring accuracy.
Summary Estimators based on Wasserstein distributionally robust optimization are obtained as solutions of min-max problems in which the statistician selects a parameter minimizing worst-case loss among all probability models within certain distance from underlying empirical measure sense. While motivated by need to identify optimal model parameters or decision choices that misspecification, these estimators recover wide range regularized estimators, including square-root lasso and support...
Motivated by applications in queueing theory, we consider a stochastic control problem whose state space is the d-dimensional positive orthant. The controlled process Z evolves as reflected Brownian motion covariance matrix exogenously specified, are its directions of reflection from orthant’s boundary surfaces. A system manager chooses drift vector [Formula: see text] at each time t based on history Z, and cost rate depends both text]. In our initial formulation, objective to minimize...
Policy learning using historical observational data are an important problem that has widespread applications. Examples include selecting offers, prices, or advertisements for consumers; choosing bids in contextual first-price auctions; and medication based on patients’ characteristics. However, existing literature rests the crucial assumption future environment where learned policy will be deployed is same as past generated data: often false too coarse approximation. In this paper, we lift...
This study investigates the effects of carbon disulfide (CS2) extraction on nanopores anthracite samples. Comprehensive pore measurement techniques, including high-pressure mercury intrusion (HPMI), low-temperature nitrogen adsorption (LTNA), and CO2 (LTCA), were employed to analyze variations nanopore structure fractal characteristics. After CS2 extraction, macropore average diameter volume increase, while area dimension decrease, indicating pore-enlarged dimension-reduced macropores. The...
Two-sided platforms are central to modern commerce and content sharing often utilize A/B testing for developing new features. While user-side experiments common, seller-side become crucial specific interventions metrics. This paper investigates the effects of interference caused by feedback loops on in two-sided platforms, with a particular focus counterfactual interleaving design, proposed \citet{ha2020counterfactual,nandy2021b}. These loops, generated pacing algorithms, cause outcomes from...
Wasserstein distributionally robust optimization estimators are obtained as solutions of min-max problems in which the statistician selects a parameter minimizing worst-case loss among all probability models within certain distance (in sense) from underlying empirical measure. While motivated by need to identify optimal model parameters or decision choices that misspecification, these recover wide range regularized estimators, including square-root lasso and support vector machines, others,...
Closed-form likelihood expansion is an important method for econometric assessment of continuous-time models driven by stochastic differential equations based on discretely sampled data. However, practical applications sophisticated usually involve significant computational efforts in calculating high-order terms order to obtain the desirable level accuracy. We provide new and efficient algorithms symbolically implementing closed-form transition density. First, combinatorial analysis leads...
We consider a reinforcement learning setting in which the deployment environment is different from training environment. Applying robust Markov decision processes formulation, we extend distributionally $Q$-learning framework studied Liu et al. [2022]. Further, improve design and analysis of their multi-level Monte Carlo estimator. Assuming access to simulator, prove that worst-case expected sample complexity our algorithm learn optimal $Q$-function within an $\epsilon$ error sup norm upper...
The digital advertising industry heavily relies on online auctions, which are mostly of first-price type. For the success a good bidding algorithm crucially accurately estimating highest bid distribution based historical data is often censored. In practice, sequence auctions takes place through multiple layers, feature that has been ignored in literature data-driven optimal strategies. this paper, we introduce two-step algorithmic procedure specifically for multi-layer auction structure....
In modern recommendation systems, the standard pipeline involves training machine learning models on historical data to predict user behaviors and improve recommendations continuously. However, these loops can introduce interference in A/B tests, where generated by control treatment algorithms, potentially with different distributions, are combined. To address challenges, we a novel approach called weighted training. This entails model probability of each point appearing either or...
We propose and study an asymptotically optimal Monte Carlo estimator for steady-state expectations of a d-dimensional reflected Brownian motion (RBM). Our is in the sense that it requires [Formula: see text] (up to logarithmic factors d) independent identically distributed scalar Gaussian random variables order output estimate with controlled error. construction based on analysis suitable multilevel strategy which, we believe, can be applied widely. This first algorithm linear complexity...