A5080575294- Statistical Methods and Inference
- Stochastic Gradient Optimization Techniques
- Privacy-Preserving Technologies in Data
- Sparse and Compressive Sensing Techniques
- Domain Adaptation and Few-Shot Learning
- Statistical Methods in Clinical Trials
- Topic Modeling
- Neural Networks and Applications
- Cryptography and Data Security
- Advanced Neural Network Applications
- Nonlinear Optical Materials Studies
- Natural Language Processing Techniques
- Adversarial Robustness in Machine Learning
- Advanced Bandit Algorithms Research
- Dermatologic Treatments and Research
- Optimal Experimental Design Methods
- Model Reduction and Neural Networks
- Porphyrin and Phthalocyanine Chemistry
- Photochemistry and Electron Transfer Studies
- Reconstructive Surgery and Microvascular Techniques
- Injection Molding Process and Properties
- Gaussian Processes and Bayesian Inference
- Gene expression and cancer classification
- Multimodal Machine Learning Applications
- Statistical Methods and Bayesian Inference
Shanghai Jiao Tong University
2014-2025
California University of Pennsylvania
2018-2025
Shanghai Ninth People's Hospital
2014-2025
Union Hospital
2021-2025
Huazhong University of Science and Technology
2021-2025
University of Pennsylvania
2015-2024
University of Science and Technology of China
2019-2024
Sun Yat-sen University
2021-2023
Feng Chia University
2017-2023
Beijing Academy of Artificial Intelligence
2023
DETR has been recently proposed to eliminate the need for many hand-designed components in object detection while demonstrating good performance. However, it suffers from slow convergence and limited feature spatial resolution, due limitation of Transformer attention modules processing image maps. To mitigate these issues, we Deformable DETR, whose only attend a small set key sampling points around reference. can achieve better performance than (especially on objects) with 10 times less...
We introduce a new pre-trainable generic representation for visual-linguistic tasks, called Visual-Linguistic BERT (VL-BERT short). VL-BERT adopts the simple yet powerful Transformer model as backbone, and extends it to take both visual linguistic embedded features input. In it, each element of input is either word from sentence, or region-of-interest (RoI) image. It designed fit most downstream tasks. To better exploit representation, we pre-train on massive-scale Conceptual Captions...
We introduce a new estimator for the vector of coefficients $\beta$ in linear model $y=X\beta+z$, where $X$ has dimensions $n\times p$ with $p$ possibly larger than $n$. SLOPE, short Sorted L-One Penalized Estimation, is solution to \[\min_{b\in\mathbb{R}^p}\frac{1}{2}\Vert y-Xb\Vert _{\ell_2}^2+\lambda_1\vert b\vert _{(1)}+\lambda_2\vert b\vert_{(2)}+\cdots+\lambda_p\vert b\vert_{(p)},\] $\lambda_1\ge\lambda_2\ge\cdots\ge\lambda_p\ge0$ and $\vert b\vert_{(1)}\ge\vert...
We derive a second-order ordinary differential equation (ODE) which is the limit of Nesterov's accelerated gradient method. This ODE exhibits approximate equivalence to scheme and thus can serve as tool for analysis. show that continuous time allows better understanding scheme. As byproduct, we obtain family schemes with similar convergence rates. The interpretation also suggests restarting leading an algorithm, be rigorously proven converge at linear rate whenever objective strongly convex.
Abstract Cutaneous wound healing, especially diabetic is a common clinical challenge. Reactive oxygen species (ROS) and bacterial infection are two major detrimental states that induce oxidative stress inflammatory responses impede angiogenesis healing. A derivative of the metabolite itaconate, 4‐octyl itaconate (4OI) has attracted increasing research interest in recent years due to its antioxidant anti‐inflammatory properties. In this study, 4OI‐modified black phosphorus (BP) nanosheets...
Abstract In the past decade, differential privacy has seen remarkable success as a rigorous and practical formalization of data privacy. This definition its divergence based relaxations, however, have several acknowledged weaknesses, either in handling composition private algorithms or analysing important primitives like amplification by subsampling. Inspired hypothesis testing formulation privacy, this paper proposes new relaxation which we term ‘f-differential privacy’ (f-DP). notion...
Repairing infected bone defects is a challenge in the field of orthopedics because limited self-healing capacity tissue and susceptibility refractory materials to bacterial activity. Innervation initiating factor for regeneration plays key regulatory role subsequent vascularization, ossification, mineralization processes. Infection leads necrosis local nerve fibers, impeding repair defects. Herein, biomaterial that can induce skeletal-associated neural network reconstruction with high...
Self-supervised learning (SSL) has delivered superior performance on a variety of downstream vision tasks. Two main-stream SSL frameworks have been proposed, i.e., Instance Discrimination (ID) and Masked Image Modeling (MIM). ID pulls together representations from different views the same image, while avoiding feature collapse. It lacks spatial sensitivity, which requires modeling local structure within each image. On other hand, MIM reconstructs original content given masked instead does...
In regression settings where explanatory variables have very low correlations and there are relatively few effects, each of large magnitude, we expect the Lasso to find important with errors, if any. This paper shows that in a regime linear sparsity—meaning fraction nonvanishing effect tends constant, however small—this cannot really be case, even when design stochastically independent. We demonstrate true features null always interspersed on path, this phenomenon occurs no matter how strong...
We consider high-dimensional sparse regression problems in which we observe $\mathbf{y}=\mathbf{X}\boldsymbol{\beta} +\mathbf{z}$, where $\mathbf{X}$ is an $n\times p$ design matrix and $\mathbf{z}$ $n$-dimensional vector of independent Gaussian errors, each with variance $\sigma^{2}$. Our focus on the recently introduced SLOPE estimator [Ann. Appl. Stat. 9 (2015) 1103–1140], regularizes least-squares estimates rank-dependent penalty $\sum_{1\le i\le p}\lambda_{i}|\widehat{\beta} |_{(i)}$,...
Differential privacy has seen remarkable success as a rigorous and practical formalization of data in the past decade. This definition its divergence based relaxations, however, have several acknowledged weaknesses, either handling composition private algorithms or analyzing important primitives like amplification by subsampling. Inspired hypothesis testing formulation privacy, this paper proposes new relaxation, which we term `$f$-differential privacy' ($f$-DP). notion number appealing...
Gradient-based optimization algorithms can be studied from the perspective of limiting ordinary differential equations (ODEs). Motivated by fact that existing ODEs do not distinguish between two fundamentally different algorithms---Nesterov's accelerated gradient method for strongly convex functions (NAG-SC) and Polyak's heavy-ball method---we study an alternative process yields high-resolution ODEs. We show these permit a general Lyapunov function framework analysis convergence in both...
Abstract Gradient-based optimization algorithms can be studied from the perspective of limiting ordinary differential equations (ODEs). Motivated by fact that existing ODEs do not distinguish between two fundamentally different algorithms—Nesterov’s accelerated gradient method for strongly convex functions (NAG-) and Polyak’s heavy-ball method—we study an alternative process yields high-resolution . We show these permit a general Lyapunov function framework analysis convergence in both...
To effectively exploit the potential of large-scale models, various pre-training strategies supported by massive data from different sources are proposed, including supervised pre-training, weakly-supervised and self-supervised pre-training. It has been proved that combining multiple modalities/sources can greatly boost training models. However, current works adopt a multi-stage system, where complex pipeline may increase uncertainty instability is thus desirable these be integrated in...
The rapid advances of large language models (LLMs), such as ChatGPT, are revolutionizing data science and statistics. These state-of-the-art tools can streamline complex processes cleaning, model building, interpretation, report writing. As a result, it reshapes the role scientists. We argue that LLMs transforming responsibilities scientists, shifting their focus from hands-on coding, data-wrangling conducting standard analyses to assessing managing performed by these automated AIs. This...
In this article, we present a detailed review of current practices and state-of-the-art methodologies in the field differential privacy (DP), with focus advancing DP's deployment real-world applications. Key points high-level contents article were originated from discussions "Differential Privacy (DP): Challenges Towards Next Frontier," workshop held July 2022 experts industry, academia, public sector seeking answers to broad questions pertaining its implications design industry-grade...
We present results of a joint computational and experimental study for series annulated metalloporphyrins in order to establish structure−property relationships. Specifically, we have examined the effects substitution by meso-tetraphenylation, tetrabenzo tetranaphtho annulation, changing central metal from zinc (Zn) palladium (Pd). Utilizing absorption emission spectroscopy laser flash photolysis techniques, photophysical properties these porphyrins been determined. Upon addition benzo or...
We introduce a novel method for sparse regression and variable selection, which is inspired by modern ideas in multiple testing. Imagine we have observations from the linear model y = X beta + z, then suggest estimating coefficients means of new estimator called SLOPE, solution to minimize 0.5 ||y - Xb\|_2^2 lambda_1 |b|_(1) lambda_2 |b|_(2) ... lambda_p |b|_(p); here, >= λ_2 λ_p 0 |b|_(p) order statistic magnitudes b. The regularizer sorted L1 norm penalizes according their rank: higher...