arXiv.itsnick.co

an independent field guide to 23 million papers from arXiv & OpenAlex — not affiliated with arXiv.org

Author

Alexander Lin

19 papers · 2012–2025

2012: 12013: 02014: 02015: 02016: 12017: 02018: 32019: 22020: 12021: 22022: 12023: 42024: 32025: 1
2012peak 4 in 20232025

Frequent co-authors

Papers

showing 1–19 of 19 RSS

Autoregressive Knowledge Distillation through Imitation Learning

Alexander Lin, Jeremy Wohlwend, Howard Chen +1 more

cs.CL · · cited 88× · arXiv ↗ · more like this ·

The performance of autoregressive models on natural language generation tasks has dramatically improved due to the adoption of deep, self-attentive architectures. However, these gains have come at the cost of hindering inference speed, making state-of-the-art models cumbersome to deploy in real-world, time-sensitive settings. We develop a compression technique for autoregressive models that is driven by an imitation learning perspective on knowledge distillation. The algorithm is designed to address the exposure bias problem. On prototypical language generation tasks such as translation and summarization, our method consistently outperforms other distillation algorithms, such as sequence-level knowledge distillation. Student models trained with our method attain 1.4 to 4.8 BLEU/ROUGE points higher than those trained from scratch, while increasing inference speed by up to 14 times in…

A Global Collaboration to Study Intimate Partner Violence-Related Head Trauma: The ENIGMA Consortium IPV Working Group

Carrie Esopenko, Jessica Meyer, Elisabeth A. Wilde +9 more

Demography · Lincoln (University of Nebraska) · · cited 6× · DOI ↗ · more like this ·

Intimate partner violence includes psychological aggression, physical violence, sexual violence, and stalking from a current or former intimate partner. Past research suggests that exposure to intimate partner violence can impact cognitive and psychological functioning, as well as neurological outcomes. These seem to be compounded in those who suffer a brain injury as a result of trauma to the head, neck or body due to physical and/or sexual violence. However, our understanding of the neurobehavioral and neurobiological effects of head trauma in this population is limited due to factors including difficulty in accessing/recruiting participants, heterogeneity of samples, and premorbid and comorbid factors that impact outcomes. Thus, the goal of the Enhancing Neuroimaging and Genetics through Meta-Analysis (ENIGMA) Consortium Intimate Partner Violence Working Group is to develop a global…

Preparing for the Plastic Surgery In-Service Exam: Evidence-Based Essentials for the Plastic Surgery Resident

Gianfranco Frojo, Kashyap Komarraju Tadisina, Sumesh Kaswan +2 more

Gender Studies · Plastic & Reconstructive Surgery · · cited 6× · DOI ↗ · more like this ·

Sir: The Plastic Surgery In-Service Training Exam is designed to objectively assess plastic surgery knowledge in five competency sections: core surgical principles, comprehensive, hand/extremities, cosmetic/breast, and craniomaxillofacial. The American Society of Plastic Surgeons In-Service Exam Committee consists of five subcommittees that oversee the creation of exam questions. The National Board of Medical Examiners administers the exam and, in combination with the In-Service Exam Committee, analyzes the performances of each question and examinee annually. Scoring at a lower percentile compared to peers on the Plastic Surgery In-Service Training Exam correlates with a higher risk of written examination failure.1 The transparency of syllabi has led to multiple publications analyzing the sources of citations and the frequency of tested subjects—specifically, lower extremity, aesthetic…

The Plastic Surgery Match

Gianfranco Frojo, Kashyap Komarraju Tadisina, Zachary Pressman +3 more

Gender Studies · Annals of Plastic Surgery · · cited 6× · DOI ↗ · more like this ·

BACKGROUND: The integrated plastic surgery match is a competitive process not only for applicants but also for programs vying for highly qualified candidates. Interactions between applicants and program constituents are limited to a single interview visit. The authors aimed to identify components of the interview visit that influence applicant decision making when determining a final program rank list. METHODS: Thirty-six applicants who were interviewed (100% response) completed the survey. Applicants rated the importance of 20 elements of the interview visit regarding future ranking of the program on a 1 to 5 Likert scale. Data were analyzed using descriptive statistics, hierarchical cluster analysis, analysis of variance, and Pearson correlations. A literature review was performed regarding the plastic surgery integrated residency interview process. RESULTS: Survey questions were…

“Secular Africa?” Making Sense of Noncompliance to Secular Constitutions in Sub-Saharan Africa

Bethania Michael, Alexander Lin, Jacques Berlinerblau

Sociology and Political Science · Journal of Church and State · · cited 5× · DOI ↗ · more like this ·

Aside from the work of Africanists, it is not widely understood that a “secular Africa” exists, or, more precisely, a sub-Saharan Africa of twenty-seven constitutionally secular states. Although an immense theoretical body of literature is devoted to “secular studies,” little of it takes these twenty-seven cases into consideration. In order to build a bridge between the findings of Africanists and students of political secularism, we advance a twelve-point typology of the constitutionally secular state. From there, we call attention to a discrepancy between the secular precepts found in African constitutions and the actual secularity practiced by governments and citizens. Using research on state fragility, postcolonial theory, and constitutionalism, we offer a few possible explanations for the “compliance gap” present in many African secular states. Finally, we reexamine fundamental…

Mixture Model Auto-Encoders: Deep Clustering through Dictionary Learning

Alexander Lin, Andrew H. Song, Demba Ba

cs.LG · · cited 5× · arXiv ↗ · more like this ·

State-of-the-art approaches for clustering high-dimensional data utilize deep auto-encoder architectures. Many of these networks require a large number of parameters and suffer from a lack of interpretability, due to the black-box nature of the auto-encoders. We introduce Mixture Model Auto-Encoders (MixMate), a novel architecture that clusters data by performing inference on a generative model. Derived from the perspective of sparse dictionary learning and mixture models, MixMate comprises several auto-encoders, each tasked with reconstructing data in a distinct cluster, while enforcing sparsity in the latent space. Through experiments on various image datasets, we show that MixMate achieves competitive performance compared to state-of-the-art deep clustering algorithms, while using orders of magnitude fewer parameters.

The Study of Intimate Partner Violence-Related Head Trauma: Recommendations from the Enhancing Neuroimaging and Genetics through Meta-Analysis Consortium Intimate Partner Violence Working Group

Carrie Esopenko, Jessica Meyer, Elisabeth A. Wilde +9 more

Sociology and Political Science · PsyArXiv (OSF Preprints) · · cited 4× · DOI ↗ · more like this ·

Intimate partner violence includes psychological aggression, physical violence, sexual violence, and stalking from a current or former intimate partner. Experiencing intimate partner violence is associated with impaired neurocognitive and psychosocial functioning, mental illness, as well as structural brain alterations. These impairments seem to be compounded by exposure to physical trauma to the head. Importantly, up to 90% of women exposed to intimate partner violence also experience some form of head trauma or even repetitive head trauma. However, research on this topic is sparse, and the neurobehavioral and neurobiological effects of head trauma in this population have not been systematically investigated. A key aim of the EnhancingNeuroimaging and Genetics through Meta-Analysis Consortium Intimate Partner Violence Working Group is to provide recommendations for the harmonization of…

Clustering Time Series with Nonlinear Dynamics: A Bayesian Non-Parametric and Particle-Based Approach

Alexander Lin, Yingzhuo Zhang, Jeremy Heng +4 more

stat.ML · · cited 3× · arXiv ↗ · more like this ·

We propose a general statistical framework for clustering multiple time series that exhibit nonlinear dynamics into an a-priori-unknown number of sub-groups. Our motivation comes from neuroscience, where an important problem is to identify, within a large assembly of neurons, subsets that respond similarly to a stimulus or contingency. Upon modeling the multiple time series as the output of a Dirichlet process mixture of nonlinear state-space models, we derive a Metropolis-within-Gibbs algorithm for full Bayesian inference that alternates between sampling cluster assignments and sampling parameter values that form the basis of the clustering. The Metropolis step employs recent innovations in particle-based methods. We apply the framework to clustering time series acquired from the prefrontal cortex of mice in an experiment designed to characterize the neural underpinnings of fear.

Word-Level Explanations for Analyzing Bias in Text-to-Image Models

Alexander Lin, Lucas Monteiro Paes, Sree Harsha Tanneru +2 more

cs.CL · · cited 2× · arXiv ↗ · more like this ·

Text-to-image models take a sentence (i.e., prompt) and generate images associated with this input prompt. These models have created award wining-art, videos, and even synthetic datasets. However, text-to-image (T2I) models can generate images that underrepresent minorities based on race and sex. This paper investigates which word in the input prompt is responsible for bias in generated images. We introduce a method for computing scores for each word in the prompt; these scores represent its influence on biases in the model's output. Our method follows the principle of \emph{explaining by removing}, leveraging masked language models to calculate the influence scores. We perform experiments on Stable Diffusion to demonstrate that our method identifies the replication of societal stereotypes in generated images.

Concentration and the Second Stage of Labor: Outcomes Associated with the Interactive Metronome

Meghan McGowan, Robin J. Ou-Yang, Markus Zei +2 more

Gender Studies · American Journal of Perinatology · · cited 2× · DOI ↗ · more like this ·

OBJECTIVE: To analyze the association between concentration, as measured by the Interactive Metronome, and a prolonged second stage of labor in nulliparous patients. STUDY DESIGN: From September 2008 to November 2009, nulliparous women at ≥34 weeks' gestation who were planning to use an epidural were asked to perform a 1-minute Interactive Metronome clapping test. Scores and demographic information were recorded. Data were then abstracted regarding each patient's labor course. The main outcome measure was the frequency of the second stage of labor exceeding 2 hours. Only patients with epidural anesthesia who completed the second stage of labor and did not require operative delivery performed for fetal indications prior to 2 full hours of pushing were included. RESULTS: Of the patients whose Interactive Metronome test scores were in the last quartile, which we associated with poor…

Toward sustainable construction 3D printing: limestone and non-calcined recycled marine clay as partial cement replacement

Harn Wei Kua, Anqi Shi, Vignesh Kajandran +9 more

Artificial Intelligence · Cleaner Engineering and Technology · · cited 1× · DOI ↗ · more like this ·

This study investigates the effects of partial substitution of Ordinary Portland Cement with different proportions of non-calcined marine clay and different grades (by particle size) of Ground Limestone to formulate 3D-printable concrete. Non-calcined clay was used because of its lower energy requirement than calcined clay. The rheology of the mixes was evaluated by extrudability, tack, Large Amplitude Oscillatory shear, and Logarithmic Stress Ramp tests. Their mechanical performance was evaluated with respect to compressive, splitting, shear, and three-point bending (flexural) strengths. Results show that adding only about 1 % of oven-dried (low temperature) marine clay and 19 % of limestone improve cohesion and build-up of static yield strength while ensuring extrudability. This study hopes to pave the way for more studies on using non-calcined marine clay as a more sustainable option…

Artifact-Minimized High-Ratio Image Compression with Preserved Analysis Fidelity

Bin Duan, Logan A. Walker, Bin Xie +4 more

Computer Vision and Pattern Recognition · bioRxiv (Cold Spring Harbor Laboratory) · · cited 1× · DOI ↗ · more like this ·

ABSTRACT Recent advances in microscopy have pushed imaging data generation to an unprecedented scale. While scientists benefit from higher spatiotemporal resolutions and larger imaging volumes, the increasing data size presents significant storage, visualization, sharing, and analysis challenges. Lossless compression typically reduces the data size by <4 fold, whereas lossy compression trades smaller data size for the loss of a precise reconstruction of the original data. Here, we develop a novel quantization method and an artifact metric for automated compression parameter optimization that preserves information fidelity. We show that, when combined with the AV1 video codec, we achieve tens to ten thousand folds of data compression while introducing negligible visual defects or quantification errors in single-molecule localization and segmentation analyses. We developed an HDF5…

Probabilistic Unrolling: Scalable, Inverse-Free Maximum Likelihood Estimation for Latent Gaussian Models

Alexander Lin, Bahareh Tolooshams, Yves Atchadé +1 more

cs.LG · · cited 1× · arXiv ↗ · more like this ·

Latent Gaussian models have a rich history in statistics and machine learning, with applications ranging from factor analysis to compressed sensing to time series analysis. The classical method for maximizing the likelihood of these models is the expectation-maximization (EM) algorithm. For problems with high-dimensional latent variables and large datasets, EM scales poorly because it needs to invert as many large covariance matrices as the number of data points. We introduce probabilistic unrolling, a method that combines Monte Carlo sampling with iterative linear solvers to circumvent matrix inversion. Our theoretical analyses reveal that unrolling and backpropagation through the iterations of the solver can accelerate gradient estimation for maximum likelihood estimation. In experiments on simulated and real data, we demonstrate that probabilistic unrolling learns latent Gaussian…

High-Dimensional Sparse Bayesian Learning without Covariance Matrices

Alexander Lin, Andrew H. Song, Berkin Bilgic +1 more

eess.SP · · cited 1× · arXiv ↗ · more like this ·

Sparse Bayesian learning (SBL) is a powerful framework for tackling the sparse coding problem. However, the most popular inference algorithms for SBL become too expensive for high-dimensional settings, due to the need to store and compute a large covariance matrix. We introduce a new inference scheme that avoids explicit construction of the covariance matrix by solving multiple linear systems in parallel to obtain the posterior moments for SBL. Our approach couples a little-known diagonal estimation result from numerical linear algebra with the conjugate gradient algorithm. On several simulations, our method scales better than existing approaches in computation time and memory, especially for structured dictionaries capable of fast matrix-vector multiplication.

LGBTQ rights in Uruguay and Secularstan

Jacques Berlinerblau, Alexander Lin, Ria Pradhan

Sociology and Political Science · · cited 0× · DOI ↗ · more like this ·

This chapter undertakes to highlight the “messy” and “complicated” nature of secularism and its associated narratives. It starts by stating a widespread assumption, namely that secularism defends LGBTQ rights while religion opposes them. The model is scrutinized in light of relevant developments in contemporary Uruguay. Upon looking at the situation on the ground in other countries, however, matters become more complex. We look at pro-LGBTQ religious movements, secular states that deploy “political homophobia,” and the influence of transnational actors on the right and on the left. The chapter ends by imagining how the ideal secular state would approach the question of LGBTQ civil rights.

Preparing for the Plastic Surgery In-Service Examination

Gianfranco Frojo, Kashyap Komarraju Tadisina, Sumesh Kaswan +2 more

Gender Studies · Plastic & Reconstructive Surgery · · cited 0× · DOI ↗ · more like this ·

For more than 70 years Plastic and Reconstructive Surgery® has been the one consistently excellent reference for every specialist who uses plastic surgery techniques or works in conjunction with a plastic surgeon. The journal brings readers up-to-the-minute reports on the latest techniques and follow-up for maxillofacial reconstruction, burn repair, cosmetic reshaping, as well as news on medicolegal issues. The cosmetic section provides expanded coverage on new procedures and techniques.

Reproducibility Made Easy: A Tool for Methodological Transparency and Efficient Standardized Reporting based on the proposed MRSinMRS Consensus

Antonia Susnjar, Antonia Kaiser, Dunja Simicic +4 more

physics.med-ph · · arXiv ↗ · more like this ·

A recent expert consensus found that non-standard reporting in MRS studies led to poor reproducibility. In order to address this, MRSinMRS guidelines were introduced; however, because of the disparate nomenclature and data formats, adoption has been slow. To get around this problem, REMY, a toolbox that supports major vendor formats, was created. By efficiently filling in important fields in the MRSinMRS table, it improves reproducibility. Even with certain hardware-related restrictions, REMY makes a substantial contribution to the completion of acquisition parameters, which facilitates reporting. Its compatibility and user-friendly interface should promote widespread adoption of MRSinMRS, raising the caliber of MRS research.

An Efficient Algorithm for Clustered Multi-Task Compressive Sensing

Alexander Lin, Demba Ba

eess.SP · · arXiv ↗ · more like this ·

This paper considers clustered multi-task compressive sensing, a hierarchical model that solves multiple compressive sensing tasks by finding clusters of tasks that leverage shared information to mutually improve signal reconstruction. The existing inference algorithm for this model is computationally expensive and does not scale well in high dimensions. The main bottleneck involves repeated matrix inversion and log-determinant computation for multiple large covariance matrices. We propose a new algorithm that substantially accelerates model inference by avoiding the need to explicitly compute these covariance matrices. Our approach combines Monte Carlo sampling with iterative linear solvers. Our experiments reveal that compared to the existing baseline, our algorithm can be up to thousands of times faster and an order of magnitude more memory-efficient.

Covariance-Free Sparse Bayesian Learning

Alexander Lin, Andrew H. Song, Berkin Bilgic +1 more

eess.SP · · arXiv ↗ · more like this ·

Sparse Bayesian learning (SBL) is a powerful framework for tackling the sparse coding problem while also providing uncertainty quantification. The most popular inference algorithms for SBL exhibit prohibitively large computational costs for high-dimensional problems due to the need to maintain a large covariance matrix. To resolve this issue, we introduce a new method for accelerating SBL inference -- named covariance-free expectation maximization (CoFEM) -- that avoids explicit computation of the covariance matrix. CoFEM solves multiple linear systems to obtain unbiased estimates of the posterior statistics needed by SBL. This is accomplished by exploiting innovations from numerical linear algebra such as preconditioned conjugate gradient and a little-known diagonal estimation rule. For a large class of compressed sensing matrices, we provide theoretical justifications for why our…