arXiv subjects

Rui Duan

Publications and source records attributed to Rui Duan.

31 records · Page 2Linked to original sources

arXiv2024

Multi-Task Learning with Summary Statistics

Multi-task learning has emerged as a powerful machine learning paradigm for integrating data from multiple sources, leveraging similarities between tasks to improve overall model performance. However, the application of multi-task learning to real-world settings is hindered by data-sharing constraints, especially in healthcare settings. To address this challenge, we propose a flexible multi-task learning framework utilizing summary statistics from various sources. Additionally, we present an adaptive parameter selection approach based on a variant of Lepski's method, allowing for data-driven tuning parameter selection when only summary statistics are available. Our systematic non-asymptotic analysis characterizes the performance of the proposed methods under various regimes of the sample complexity and overlap. We demonstrate our theoretical findings and the performance of the method through extensive simulations. This work offers a more flexible tool for training related models across various domains, with practical implications in genetic risk prediction and many other fields.

stat.ME↗

arXiv2023

Parrot-Trained Adversarial Examples: Pushing the Practicality of Black-Box Audio Attacks against Speaker Recognition Models

Audio adversarial examples (AEs) have posed significant security challenges to real-world speaker recognition systems. Most black-box attacks still require certain information from the speaker recognition model to be effective (e.g., keeping probing and requiring the knowledge of similarity scores). This work aims to push the practicality of the black-box attacks by minimizing the attacker's knowledge about a target speaker recognition model. Although it is not feasible for an attacker to succeed with completely zero knowledge, we assume that the attacker only knows a short (or a few seconds) speech sample of a target speaker. Without any probing to gain further knowledge about the target model, we propose a new mechanism, called parrot training, to generate AEs against the target model. Motivated by recent advancements in voice conversion (VC), we propose to use the one short sentence knowledge to generate more synthetic speech samples that sound like the target speaker, called parrot speech. Then, we use these parrot speech samples to train a parrot-trained(PT) surrogate model for the attacker. Under a joint transferability and perception framework, we investigate different ways to generate AEs on the PT model (called PT-AEs) to ensure the PT-AEs can be generated with high transferability to a black-box target model with good human perceptual quality. Real-world experiments show that the resultant PT-AEs achieve the attack success rates of 45.8% - 80.8% against the open-source models in the digital-line scenario and 47.9% - 58.3% against smart devices, including Apple HomePod (Siri), Amazon Echo, and Google Home, in the over-the-air scenario.

cs.SD↗

arXiv2023

Robust angle-based transfer learning in high dimensions

Transfer learning aims to improve the performance of a target model by leveraging data from related source populations, which is known to be especially helpful in cases with insufficient target data. In this paper, we study the problem of how to train a high-dimensional ridge regression model using limited target data and existing regression models trained in heterogeneous source populations. We consider a practical setting where only the parameter estimates of the fitted source models are accessible, instead of the individual-level source data. Under the setting with only one source model, we propose a novel flexible angle-based transfer learning (angleTL) method, which leverages the concordance between the source and the target model parameters. We show that angleTL unifies several benchmark methods by construction, including the target-only model trained using target data alone, the source model fitted on source data, and distance-based transfer learning method that incorporates the source parameter estimates and the target data under a distance-based similarity constraint. We also provide algorithms to effectively incorporate multiple source models accounting for the fact that some source models may be more helpful than others. Our high-dimensional asymptotic analysis provides interpretations and insights regarding when a source model can be helpful to the target model, and demonstrates the superiority of angleTL over other benchmark methods. We perform extensive simulation studies to validate our theoretical conclusions and show the feasibility of applying angleTL to transfer existing genetic risk prediction models across multiple biobanks.

stat.ME↗

arXiv2023

Federated Adaptive Causal Estimation (FACE) of Target Treatment Effects

Federated learning of causal estimands may greatly improve estimation efficiency by leveraging data from multiple study sites, but robustness to heterogeneity and model misspecifications is vital for ensuring validity. We develop a Federated Adaptive Causal Estimation (FACE) framework to incorporate heterogeneous data from multiple sites to provide treatment effect estimation and inference for a flexibly specified target population of interest. FACE accounts for site-level heterogeneity in the distribution of covariates through density ratio weighting. To safely incorporate source sites and avoid negative transfer, we introduce an adaptive weighting procedure via a penalized regression, which achieves both consistency and optimal efficiency. Our strategy is communication-efficient and privacy-preserving, allowing participating sites to share summary statistics only once with other sites. We conduct both theoretical and numerical evaluations of FACE and apply it to conduct a comparative effectiveness study of BNT162b2 (Pfizer) and mRNA-1273 (Moderna) vaccines on COVID-19 outcomes in U.S. veterans using electronic health records from five VA regional sites. We show that compared to traditional methods, FACE meaningfully increases the precision of treatment effect estimates, with reductions in standard errors ranging from $26\%$ to $67\%$.

stat.ME↗

arXiv2022

Perception-Aware Attack: Creating Adversarial Music via Reverse-Engineering Human Perception

Recently, adversarial machine learning attacks have posed serious security threats against practical audio signal classification systems, including speech recognition, speaker recognition, and music copyright detection. Previous studies have mainly focused on ensuring the effectiveness of attacking an audio signal classifier via creating a small noise-like perturbation on the original signal. It is still unclear if an attacker is able to create audio signal perturbations that can be well perceived by human beings in addition to its attack effectiveness. This is particularly important for music signals as they are carefully crafted with human-enjoyable audio characteristics. In this work, we formulate the adversarial attack against music signals as a new perception-aware attack framework, which integrates human study into adversarial attack design. Specifically, we conduct a human study to quantify the human perception with respect to a change of a music signal. We invite human participants to rate their perceived deviation based on pairs of original and perturbed music signals, and reverse-engineer the human perception process by regression analysis to predict the human-perceived deviation given a perturbed signal. The perception-aware attack is then formulated as an optimization problem that finds an optimal perturbation signal to minimize the prediction of perceived deviation from the regressed human perception model. We use the perception-aware framework to design a realistic adversarial music attack against YouTube's copyright detector. Experiments show that the perception-aware attack produces adversarial music with significantly better perceptual quality than prior work.

cs.SD↗

arXiv2022

Generalized Federated Learning via Sharpness Aware Minimization

Federated Learning (FL) is a promising framework for performing privacy-preserving, distributed learning with a set of clients. However, the data distribution among clients often exhibits non-IID, i.e., distribution shift, which makes efficient optimization difficult. To tackle this problem, many FL algorithms focus on mitigating the effects of data heterogeneity across clients by increasing the performance of the global model. However, almost all algorithms leverage Empirical Risk Minimization (ERM) to be the local optimizer, which is easy to make the global model fall into a sharp valley and increase a large deviation of parts of local clients. Therefore, in this paper, we revisit the solutions to the distribution shift problem in FL with a focus on local learning generality. To this end, we propose a general, effective algorithm, \texttt{FedSAM}, based on Sharpness Aware Minimization (SAM) local optimizer, and develop a momentum FL algorithm to bridge local and global models, \texttt{MoFedSAM}. Theoretically, we show the convergence analysis of these two algorithms and demonstrate the generalization bound of \texttt{FedSAM}. Empirically, our proposed algorithms substantially outperform existing FL studies and significantly decrease the learning deviation.

cs.LG↗

arXiv2022

SODA: Site Object Detection dAtaset for Deep Learning in Construction

Computer vision-based deep learning object detection algorithms have been developed sufficiently powerful to support the ability to recognize various objects. Although there are currently general datasets for object detection, there is still a lack of large-scale, open-source dataset for the construction industry, which limits the developments of object detection algorithms as they tend to be data-hungry. Therefore, this paper develops a new large-scale image dataset specifically collected and annotated for the construction site, called Site Object Detection dAtaset (SODA), which contains 15 kinds of object classes categorized by workers, materials, machines, and layout. Firstly, more than 20,000 images were collected from multiple construction sites in different site conditions, weather conditions, and construction phases, which covered different angles and perspectives. After careful screening and processing, 19,846 images including 286,201 objects were then obtained and annotated with labels in accordance with predefined categories. Statistical analysis shows that the developed dataset is advantageous in terms of diversity and volume. Further evaluation with two widely-adopted object detection algorithms based on deep learning (YOLO v3/ YOLO v4) also illustrates the feasibility of the dataset for typical construction scenarios, achieving a maximum mAP of 81.47%. In this manner, this research contributes a large-scale image dataset for the development of deep learning-based object detection methods in the construction industry and sets up a performance benchmark for further evaluation of corresponding algorithms in this area.

cs.CV↗

arXiv2021

Context-Aware Online Client Selection for Hierarchical Federated Learning

Federated Learning (FL) has been considered as an appealing framework to tackle data privacy issues of mobile devices compared to conventional Machine Learning (ML). Using Edge Servers (ESs) as intermediaries to perform model aggregation in proximity can reduce the transmission overhead, and it enables great potentials in low-latency FL, where the hierarchical architecture of FL (HFL) has been attracted more attention. Designing a proper client selection policy can significantly improve training performance, and it has been extensively used in FL studies. However, to the best of our knowledge, there are no studies focusing on HFL. In addition, client selection for HFL faces more challenges than conventional FL, e.g., the time-varying connection of client-ES pairs and the limited budget of the Network Operator (NO). In this paper, we investigate a client selection problem for HFL, where the NO learns the number of successful participating clients to improve the training performance (i.e., select as many clients in each round) as well as under the limited budget on each ES. An online policy, called Context-aware Online Client Selection (COCS), is developed based on Contextual Combinatorial Multi-Armed Bandit (CC-MAB). COCS observes the side-information (context) of local computing and transmission of client-ES pairs and makes client selection decisions to maximize NO's utility given a limited budget. Theoretically, COCS achieves a sublinear regret compared to an Oracle policy on both strongly convex and non-convex HFL. Simulation results also support the efficiency of the proposed COCS policy on real-world datasets.

cs.LG↗

arXiv2021

Targeting Underrepresented Populations in Precision Medicine: A Federated Transfer Learning Approach

The limited representation of minorities and disadvantaged populations in large-scale clinical and genomics research has become a barrier to translating precision medicine research into practice. Due to heterogeneity across populations, risk prediction models are often found to be underperformed in these underrepresented populations, and therefore may further exacerbate known health disparities. In this paper, we propose a two-way data integration strategy that integrates heterogeneous data from diverse populations and from multiple healthcare institutions via a federated transfer learning approach. The proposed method can handle the challenging setting where sample sizes from different populations are highly unbalanced. With only a small number of communications across participating sites, the proposed method can achieve performance comparable to the pooled analysis where individual-level data are directly pooled together. We show that the proposed method improves the estimation and prediction accuracy in underrepresented populations, and reduces the gap of model performance across populations. Our theoretical analysis reveals how estimation accuracy is influenced by communication budgets, privacy restrictions, and heterogeneity across populations. We demonstrate the feasibility and validity of our methods through numerical experiments and a real application to a multi-center study, in which we construct polygenic risk prediction models for Type II diabetes in AA population.

stat.ML↗

arXiv2021

Heterogeneity-aware and communication-efficient distributed statistical inference

In multicenter research, individual-level data are often protected against sharing across sites. To overcome the barrier of data sharing, many distributed algorithms, which only require sharing aggregated information, have been developed. The existing distributed algorithms usually assume the data are homogeneously distributed across sites. This assumption ignores the important fact that the data collected at different sites may come from various sub-populations and environments, which can lead to heterogeneity in the distribution of the data. Ignoring the heterogeneity may lead to erroneous statistical inference. In this paper, we propose distributed algorithms which account for the heterogeneous distributions by allowing site-specific nuisance parameters. The proposed methods extend the surrogate likelihood approach to the heterogeneous setting by applying a novel density ratio tilting method to the efficient score function. The proposed algorithms maintain the same communication cost as the existing communication-efficient algorithms. We establish a non-asymptotic risk bound for the proposed distributed estimator and its limiting distribution in the two-index asymptotic setting which allows both sample size per site and the number of sites to go to infinity. In addition, we show that the asymptotic variance of the estimator attains the Cramér-Rao lower bound when the number of sites is in rate smaller than the sample size at each site. Finally, we use simulation studies and a real data application to demonstrate the validity and feasibility of the proposed methods.

stat.ME↗

arXiv2021

On the global identifiability of logistic regression models with misclassified outcomes

In the last decade, the secondary use of large data from health systems, such as electronic health records, has demonstrated great promise in advancing biomedical discoveries and improving clinical decision making. However, there is an increasing concern about biases in association studies caused by misclassification in the binary outcomes derived from electronic health records. We revisit the classical logistic regression model with misclassified outcomes. Despite that local identification conditions in some related settings have been previously established, the global identification of such models remains largely unknown and is an important question yet to be answered. We derive necessary and sufficient conditions for global identifiability of logistic regression models with misclassified outcomes, using a novel approach termed as the submodel analysis, and a technique adapted from the Picard-Lindelöf existence theorem in ordinary differential equations. In particular, our results are applicable to logistic models with discrete covariates, which is a common situation in biomedical studies, The conditions are easy to verify in practice. In addition to model identifiability, we propose a hypothesis testing procedure for regression coefficients in the misclassified logistic regression model when the model is not identifiable under the null.

math.ST↗

arXiv2020

Testing for publication bias in meta-analysis under Copas selection model

In meta-analyses, publication bias is a well-known, important and challenging issue because the validity of the results from a meta-analysis is threatened if the sample of studies retrieved for review is biased. One popular method to deal with publication bias is the Copas selection model, which provides a flexible sensitivity analysis for correcting the estimates with considerable insight into the data suppression mechanism. However, rigorous testing procedures under the Copas selection model to detect bias are lacking. To fill this gap, we develop a score-based test for detecting publication bias under the Copas selection model. We reveal that the behavior of the standard score test statistic is irregular because the parameters of the Copas selection model disappear under the null hypothesis, leading to an identifiability problem. We propose a novel test statistic and derive its limiting distribution. A bootstrap procedure is provided to obtain the p-value of the test for practical applications. We conduct extensive Monte Carlo simulations to evaluate the performance of the proposed test and apply the method to several existing meta-analyses.

stat.ME↗

arXiv2020

Missing at Random or Not: A Semiparametric Testing Approach

Practical problems with missing data are common, and statistical methods have been developed concerning the validity and/or efficiency of statistical procedures. On a central focus, there have been longstanding interests on the mechanism governing data missingness, and correctly deciding the appropriate mechanism is crucially relevant for conducting proper practical investigations. The conventional notions include the three common potential classes -- missing completely at random, missing at random, and missing not at random. In this paper, we present a new hypothesis testing approach for deciding between missing at random and missing not at random. Since the potential alternatives of missing at random are broad, we focus our investigation on a general class of models with instrumental variables for data missing not at random. Our setting is broadly applicable, thanks to that the model concerning the missing data is nonparametric, requiring no explicit model specification for the data missingness. The foundational idea is to develop appropriate discrepancy measures between estimators whose properties significantly differ only when missing at random does not hold. We show that our new hypothesis testing approach achieves an objective data oriented choice between missing at random or not. We demonstrate the feasibility, validity, and efficacy of the new test by theoretical analysis, simulation studies, and a real data analysis.

stat.ME↗