Xiangyu Chang
YOU?
Author Swipe
View article: Structural evaluation of cracked shield tunnels using computer-vision-based model updating techniques
Structural evaluation of cracked shield tunnels using computer-vision-based model updating techniques Open
View article: Can a One-Point Feedback Zeroth-order Algorithm Achieve Linear Dimension Dependent Sample Complexity?
Can a One-Point Feedback Zeroth-order Algorithm Achieve Linear Dimension Dependent Sample Complexity? Open
We revisit the one-point feedback zeroth-order (ZO) optimization problem, a classical setting in derivative-free optimization where only a single noisy function evaluation is available per query. Compared to their two-point counterparts, e…
View article: When and How Unlabeled Data Provably Improve In-Context Learning
When and How Unlabeled Data Provably Improve In-Context Learning Open
Recent research shows that in-context learning (ICL) can be effective even when demonstrations have missing or incorrect labels. To shed light on this capability, we examine a canonical setting where the demonstrations are drawn according …
View article: Stochastic Diagonal Estimation Based on Matrix Quadratic Form Oracles
Stochastic Diagonal Estimation Based on Matrix Quadratic Form Oracles Open
We study the problem of estimating the diagonal of an implicitly given matrix $\Ab$. For such a matrix we have access to an oracle that allows us to evaluate the matrix quadratic form $ \ub^\top \Ab \ub$. Based on this query oracle, we pro…
View article: Privacy Leaks by Adversaries: Adversarial Iterations for Membership Inference Attack
Privacy Leaks by Adversaries: Adversarial Iterations for Membership Inference Attack Open
Membership inference attack (MIA) has become one of the most widely used and effective methods for evaluating the privacy risks of machine learning models. These attacks aim to determine whether a specific sample is part of the model's tra…
View article: Towards Understanding The Calibration Benefits of Sharpness-Aware Minimization
Towards Understanding The Calibration Benefits of Sharpness-Aware Minimization Open
Deep neural networks have been increasingly used in safety-critical applications such as medical diagnosis and autonomous driving. However, many studies suggest that they are prone to being poorly calibrated and have a propensity for overc…
View article: Continuum-armed Bandit Optimization with Batch Pairwise Comparison Oracles
Continuum-armed Bandit Optimization with Batch Pairwise Comparison Oracles Open
This paper studies a bandit optimization problem where the goal is to maximize a function $f(x)$ over $T$ periods for some unknown strongly concave function $f$. We consider a new pairwise comparison oracle, where the decision-maker choose…
View article: Provable Benefits of Task-Specific Prompts for In-context Learning
Provable Benefits of Task-Specific Prompts for In-context Learning Open
The in-context learning capabilities of modern language models have motivated a deeper mathematical understanding of sequence models. A line of recent work has shown that linear attention models can emulate projected gradient descent itera…
View article: Transcranial direct current stimulation and lesions hierarchically reorganize brain network dynamics with biological annotations
Transcranial direct current stimulation and lesions hierarchically reorganize brain network dynamics with biological annotations Open
View article: Randomized Spectral Clustering for Large-Scale Multi-Layer Networks
Randomized Spectral Clustering for Large-Scale Multi-Layer Networks Open
Large-scale multi-layer networks with large numbers of nodes, edges, and layers arise across various domains, which poses a great computational challenge for the downstream analysis. In this paper, we develop an efficient randomized spectr…
View article: Learned Low-Rank Representation and its Theoretical Convergence Analysis
Learned Low-Rank Representation and its Theoretical Convergence Analysis Open
View article: Selective Attention: Enhancing Transformer through Principled Context Control
Selective Attention: Enhancing Transformer through Principled Context Control Open
The attention mechanism within the transformer architecture enables the model to weigh and combine tokens based on their relevance to the query. While self-attention has enjoyed major success, it notably treats all queries $q$ in the same …
View article: Towards Data Valuation via Asymmetric Data Shapley
Towards Data Valuation via Asymmetric Data Shapley Open
As data emerges as a vital driver of technological and economic advancements, a key challenge is accurately quantifying its value in algorithmic decision-making. The Shapley value, a well-established concept from cooperative game theory, h…
View article: AdapFair: Ensuring Adaptive Fairness for Machine Learning Operations
AdapFair: Ensuring Adaptive Fairness for Machine Learning Operations Open
The biases and discrimination of machine learning algorithms have attracted significant attention, leading to the development of various algorithms tailored to specific contexts. However, these solutions often fall short of addressing fair…
View article: A new algorithm for computation of a regularization solution path for reinforced multicategory support vector machines
A new algorithm for computation of a regularization solution path for reinforced multicategory support vector machines Open
The recently proposed Reinforced Multicategory Support Vector Machine (RMSVM) has been proven to have desirable theoretical properties as well as competitive numerical accuracy for multi-class classification problems. Currently solving the…
View article: Uncertainty Quantification of Data Shapley via Statistical Inference
Uncertainty Quantification of Data Shapley via Statistical Inference Open
As data plays an increasingly pivotal role in decision-making, the emergence of data markets underscores the growing importance of data valuation. Within the machine learning landscape, Data Shapley stands out as a widely embraced method f…
View article: Double Variance Reduction: A Smoothing Trick for Composite Optimization Problems without First-Order Gradient
Double Variance Reduction: A Smoothing Trick for Composite Optimization Problems without First-Order Gradient Open
Variance reduction techniques are designed to decrease the sampling variance, thereby accelerating convergence rates of first-order (FO) and zeroth-order (ZO) optimization methods. However, in composite optimization problems, ZO methods en…
View article: Anderson Acceleration Without Restart: A Novel Method with $n$-Step Super Quadratic Convergence Rate
Anderson Acceleration Without Restart: A Novel Method with $n$-Step Super Quadratic Convergence Rate Open
In this paper, we propose a novel Anderson's acceleration method to solve nonlinear equations, which does \emph{not} require a restart strategy to achieve numerical stability. We propose the greedy and random versions of our algorithm. Spe…
View article: FLASH: Federated Learning Across Simultaneous Heterogeneities
FLASH: Federated Learning Across Simultaneous Heterogeneities Open
The key premise of federated learning (FL) is to train ML models across a diverse set of data-owners (clients), without exchanging local data. An overarching challenge to this date is client heterogeneity, which may arise not only from var…
View article: Plug-and-Play Transformer Modules for Test-Time Adaptation
Plug-and-Play Transformer Modules for Test-Time Adaptation Open
Parameter-efficient tuning (PET) methods such as LoRA, Adapter, and Visual Prompt Tuning (VPT) have found success in enabling adaptation to new domains by tuning small modules within a transformer model. However, the number of domains enco…
View article: Short-Term Passenger Flow Prediction Based on a Hybrid Method in Nanjing Metro System
Short-Term Passenger Flow Prediction Based on a Hybrid Method in Nanjing Metro System Open
View article: Optimal Decentralized Composite Optimization for Convex Functions
Optimal Decentralized Composite Optimization for Convex Functions Open
In this paper, we focus on the decentralized composite optimization for convex functions. Because of advantages such as robust to the network and no communication bottle-neck in the central server, the decentralized optimization has attrac…
View article: Prediction and Early Warning of Extreme Winds for High-Speed Railway Bridge Construction Using Machine-Learning Methods
Prediction and Early Warning of Extreme Winds for High-Speed Railway Bridge Construction Using Machine-Learning Methods Open
Measuring the impact of extreme winds is important in high-speed railway bridge construction to avoid the risk of engineering accidents. This study presents an early-warning framework for high-speed railway (HSR) bridge construction under …
View article: PPFL: A Personalized Federated Learning Framework for Heterogeneous Population
PPFL: A Personalized Federated Learning Framework for Heterogeneous Population Open
Personalization aims to characterize individual preferences and is widely applied across many fields. However, conventional personalized methods operate in a centralized manner, potentially exposing raw data when pooling individual informa…
View article: A New Causal Rule Learning Approach to Interpretable Estimation of Heterogeneous Treatment Effect
A New Causal Rule Learning Approach to Interpretable Estimation of Heterogeneous Treatment Effect Open
Interpretability plays a crucial role in the application of statistical learning to estimate heterogeneous treatment effects (HTE) in complex diseases. In this study, we leverage a rule-based workflow, namely causal rule learning (CRL), to…
View article: FedYolo: Augmenting Federated Learning with Pretrained Transformers
FedYolo: Augmenting Federated Learning with Pretrained Transformers Open
The growth and diversity of machine learning applications motivate a rethinking of learning with mobile and edge devices. How can we address diverse client goals and learn with scarce heterogeneous data? While federated learning aims to ad…
View article: Privacy-Preserving Community Detection for Locally Distributed Multiple Networks
Privacy-Preserving Community Detection for Locally Distributed Multiple Networks Open
Modern multi-layer networks are commonly stored and analyzed in a local and distributed fashion because of the privacy, ownership, and communication costs. The literature on the model-based statistical methods for community detection based…
View article: 2D-Shapley: A Framework for Fragmented Data Valuation
2D-Shapley: A Framework for Fragmented Data Valuation Open
Data valuation -- quantifying the contribution of individual data sources to certain predictive behaviors of a model -- is of great importance to enhancing the transparency of machine learning and designing incentive systems for data shari…
View article: Subsampling-Based Modified Bayesian Information Criterion for Large-Scale Stochastic Block Models
Subsampling-Based Modified Bayesian Information Criterion for Large-Scale Stochastic Block Models Open
Identifying the number of communities is a fundamental problem in community detection, which has received increasing attention recently. However, rapid advances in technology have led to the emergence of large-scale networks in various dis…
View article: Learning Personalized Brain Functional Connectivity of MDD Patients from Multiple Sites via Federated Bayesian Networks
Learning Personalized Brain Functional Connectivity of MDD Patients from Multiple Sites via Federated Bayesian Networks Open
Identifying functional connectivity biomarkers of major depressive disorder (MDD) patients is essential to advance understanding of the disorder mechanisms and early intervention. However, due to the small sample size and the high dimensio…