Search

We study the problem of finding the index of the minimum value of a vector from noisy observations. This problem is relevant in population/policy comparison, discrete maximum likelihood, and model selection. We develop an asymptotically normal test statistic, even in high-dimensional settings and with potentially many ties in the population mean vector, by integrating concepts and tools from cross-validation and differential privacy. The key technical ingredient is a central limit theorem for globally dependent data. We also propose practical ways to select the tuning parameter that adapts to the signal landscape. Numerical experiments and data examples demonstrate the ability of the proposed method to achieve a favorable bias-variance trade-off in practical scenarios.

The emergence of large language models (LLMs), and their increased use in user-facing systems, has led to substantial privacy concerns. To date, research on these privacy concerns has been model-centered: exploring how LLMs lead to privacy risks like memorization, or can be used to infer personal characteristics about people from their content. We argue that there is a need for more research focusing on the human aspect of these privacy issues: e.g., research on how design paradigms for LLMs affect users' disclosure behaviors, users' mental models and preferences for privacy controls, and the design of tools, systems, and artifacts that empower end-users to reclaim ownership over their personal data. To build usable, efficient, and privacy-friendly systems powered by these models with imperfect privacy properties, our goal is to initiate discussions to outline an agenda for conducting human-centered research on privacy issues in LLM-powered systems. This Special Interest Group (SIG) aims to bring together researchers with backgrounds in usable security and privacy, human-AI collaboration, NLP, or any other related domains to share their perspectives and experiences on this problem, to help our community establish a collective understanding of the challenges, research opportunities, research methods, and strategies to collaborate with researchers outside of HCI., Comment: 4 pages, CHI EA'24

Visualization and assessment of copula structures are crucial for accurately understanding and modeling the dependencies in multivariate data analysis. In this paper, we introduce an innovative method that employs functional boxplots and rank-based testing procedures to evaluate copula symmetry. This approach is specifically designed to assess key characteristics such as reflection symmetry, radial symmetry, and joint symmetry. We first construct test functions for each specific property and then investigate the asymptotic properties of their empirical estimators. We demonstrate that the functional boxplot of these sample test functions serves as an informative visualization tool of a given copula structure, effectively measuring the departure from zero of the test function. Furthermore, we introduce a nonparametric testing procedure to assess the significance of deviations from symmetry, ensuring the accuracy and reliability of our visualization method. Through extensive simulation studies involving various copula models, we demonstrate the effectiveness of our testing approach. Finally, we apply our visualization and testing techniques to two real-world datasets: a nutritional habits survey with five variables and wind speed data from three locations in Saudi Arabia.

Privacy is a key principle for developing ethical AI technologies, but how does including AI technologies in products and services change privacy risks? We constructed a taxonomy of AI privacy risks by analyzing 321 documented AI privacy incidents. We codified how the unique capabilities and requirements of AI technologies described in those incidents generated new privacy risks, exacerbated known ones, or otherwise did not meaningfully alter the risk. We present 12 high-level privacy risks that AI technologies either newly created (e.g., exposure risks from deepfake pornography) or exacerbated (e.g., surveillance risks from collecting training data). One upshot of our work is that incorporating AI technologies into a product can alter the privacy risks it entails. Yet, current approaches to privacy-preserving AI/ML (e.g., federated learning, differential privacy, checklists) only address a subset of the privacy risks arising from the capabilities and data requirements of AI.

The widespread use of Large Language Model (LLM)-based conversational agents (CAs), especially in high-stakes domains, raises many privacy concerns. Building ethical LLM-based CAs that respect user privacy requires an in-depth understanding of the privacy risks that concern users the most. However, existing research, primarily model-centered, does not provide insight into users' perspectives. To bridge this gap, we analyzed sensitive disclosures in real-world ChatGPT conversations and conducted semi-structured interviews with 19 LLM-based CA users. We found that users are constantly faced with trade-offs between privacy, utility, and convenience when using LLM-based CAs. However, users' erroneous mental models and the dark patterns in system design limited their awareness and comprehension of the privacy risks. Additionally, the human-like interactions encouraged more sensitive disclosures, which complicated users' ability to navigate the trade-offs. We discuss practical design guidelines and the needs for paradigm shifts to protect the privacy of LLM-based CA users., Comment: 26 pages, 5 figures

Let $K / \mathbb{Q}_p$ be a finite unramified extension, and let $\mathcal{X}_n$ denote the Emerton-Gee stack parametrizing \'etale $(\varphi,\Gamma_K)$-modules of rank $n$. It is known since the work of Emerton-Gee that the irreducible components of the reduced special fiber of $\mathcal{X}$ are labeled by Serre weights $\sigma$ of $\operatorname{GL}_n(k)$. If such a component is denoted $\mathcal{X}(\sigma)$, we prove that $\mathcal{O}(\mathcal{X}(\sigma)) \cong \mathbb{F}[x_1,x_2,\dots,x_{n-1},x_n^{\pm 1}]$ when $\sigma$ is sufficiently generic., Comment: Corrected an error in the proof of Theorem A.3.7 and made numerous minor changes

Let $G$ be a $p$-adic Lie group associated to a connected reductive group over $\mathbb{Q}_{p}$. Let $P$ be a parabolic subgroup of $G$ and let $M$ be a Levi quotient of $P$. In this paper, we define a $\delta$-functor $H^{\star}J_{P}$ from the category of admissible locally analytic $G$-representations to the category of essentially admissible locally analytic $M$-representations that extends the Jacquet-Emerton module functor $J_{P}$ defined by Emerton., Comment: Any comments are welcomed!

Deriving disease subtypes from electronic health records (EHRs) can guide next-generation personalized medicine. However, challenges in summarizing and representing patient data prevent widespread practice of scalable EHR-based stratification analysis. Here we present an unsupervised framework based on deep learning to process heterogeneous EHRs and derive patient representations that can efficiently and effectively enable patient stratification at scale. We considered EHRs of 1,608,741 patients from a diverse hospital cohort comprising of a total of 57,464 clinical concepts. We introduce a representation learning model based on word embeddings, convolutional neural networks, and autoencoders (i.e., ConvAE) to transform patient trajectories into low-dimensional latent vectors. We evaluated these representations as broadly enabling patient stratification by applying hierarchical clustering to different multi-disease and disease-specific patient cohorts. ConvAE significantly outperformed several baselines in a clustering task to identify patients with different complex conditions, with 2.61 entropy and 0.31 purity average scores. When applied to stratify patients within a certain condition, ConvAE led to various clinically relevant subtypes for different disorders, including type 2 diabetes, Parkinson's disease and Alzheimer's disease, largely related to comorbidities, disease progression, and symptom severity. With these results, we demonstrate that ConvAE can generate patient representations that lead to clinically meaningful insights. This scalable framework can help better understand varying etiologies in heterogeneous sub-populations and unlock patterns for EHR-based research in the realm of personalized medicine., Comment: C.F. and R.M. share senior authorship

Constructing gene regulatory networks is a critical step in revealing disease mechanisms from transcriptomic data. In this work, we present NO-BEARS, a novel algorithm for estimating gene regulatory networks. The NO-BEARS algorithm is built on the basis of the NOTEARS algorithm with two improvements. First, we propose a new constraint and its fast approximation to reduce the computational cost of the NO-TEARS algorithm. Next, we introduce a polynomial regression loss to handle non-linearity in gene expressions. Our implementation utilizes modern GPU computation that can decrease the time of hours-long CPU computation to seconds. Using synthetic data, we demonstrate improved performance, both in processing time and accuracy, on inferring gene regulatory networks from gene expression data., Comment: Preprint of an article submitted for consideration in Pacific Symposium on Biocomputing copyright 2019 World Scientific Publishing Co., Singapore, http://psb.stanford.edu/http://psb.stanford.edu/

Given the crucial role of microtubules for cell survival, many researchers have found success using microtubule-targeting agents in the search for effective cancer therapeutics. Understanding microtubule responses to targeted interventions requires that the microtubule network within cells can be consistently observed across a large sample of images. However, fluorescence noise sources captured simultaneously with biological signals while using wide-field microscopes can obfuscate fine microtubule structures. Such requirements are particularly challenging for high-throughput imaging, where researchers must make decisions related to the trade-off between imaging quality and speed. Here, we propose a computational framework to enhance the quality of high-throughput imaging data to achieve fast speed and high quality simultaneously. Using CycleGAN, we learn an image model from low-throughput, high-resolution images to enhance features, such as microtubule networks in high-throughput low-resolution images. We show that CycleGAN is effective in identifying microtubules with 0.93+ AUC-ROC and that these results are robust to different kinds of image noise. We further apply CycleGAN to quantify the changes in microtubule density as a result of the application of drug compounds, and show that the quantified responses correspond well with known drug effects, Comment: accepted and presented in Machine Learning for Healthcare 2019

Transition metal dichalcogenides (TMDs) are layered semiconducting van der Waal crystals and promising materials for a wide range of electronic and optoelectronic devices. Realizing practical electrical and optoelectronic device applications requires a connection between a metal junction and a TMD semiconductor. Hence, a complete understanding of electronic band alignments and the potential barrier heights governing the transport through a metal-TMD-metal junction is critical. But, there is a knowledge gap; it is not clear how the energy bands of a TMD align while in contact with a metal as a function of the number of layers. In pursuit of removing this knowledge gap, we have performed conductive atomic force microscopy (CAFM) of few layered (1-5) MoS2 immobilized on ultra-flat conducting Au surfaces (root mean square (RMS) surface roughness <0.2 nm) and indium tin oxide (ITO) substrate (RMS surface roughness <0.7 nm) forming a vertical metal (conductive-AFM tip)-semiconductor-metal device. We have observed that the current increases as the number of layers increases up to 5 layers. By applying Fowler-Nordheim tunneling theory, we have determined the barrier heights for different layers and observed that the barrier height decreases as the number of layers increases. Using density functional theory (DFT) calculation, we successfully demonstrated that the barrier height decreases as the layer number increases. By illuminating the TMDs on a transparent ultra-flat conducting ITO substrate, we observed a reduction in current when compared to the current measured in the dark, hence demonstrating negative photoconductivity. Our study provides a fundamental understanding of the local electronic and optoelectronic behaviors of TMD-metal junction, and may pave an avenue toward developing nanoscale electronic devices with tailored layer-dependent transport properties., Comment: 11 pages, 5 figures