Search

This paper explores opportunities and challenges for data-driven advocacy to support home care workers, an often overlooked group of low-wage, frontline health workers. First, we investigate what data to collect and how to collect it in ways that preserve privacy and avoid burdening workers. Second, we examine how workers and advocates could use collected data to strengthen individual and collective advocacy efforts. Our qualitative study with 11 workers and 15 advocates highlights tensions between workers' desires for individual and immediate benefits and advocates' preferences to prioritize more collective and long-term benefits. We also uncover discrepancies between participants' expectations for how data might transform advocacy and their on-the-ground experiences collecting and using real data. Finally, we discuss future directions for data-driven worker advocacy, including combining different kinds of data to ameliorate challenges, leveraging advocates as data stewards, and accounting for workers' and organizations' heterogeneous goals., Comment: Accepted to CHI 2025

Researchers increasingly look to understand experiences of pain, harm, and marginalization via qualitative analysis. Such work is needed to understand and address social ills, but poses risks to researchers' well-being: sifting through volumes of data on painful human experiences risks incurring traumatic exposure in the researcher. In this paper, we explore how the principles of trauma-informed computing (TIC) can be applied to reimagine healthier tools and workflows for qualitative analysis. We apply TIC to create a design provocation called TIQA, a system for qualitative coding that leverages language modeling, semantic search, and recommendation systems to measure and mitigate an analyst's exposure to concepts they find traumatic. Through a formative study of TIQA with 15 participants, we illuminate the complexities of enacting TIC in qualitative knowledge infrastructure, and potential roles for machine assistance in mitigating researchers' trauma. To assist scholars in translating the high-level principles of TIC into sociotechnical system design, we argue for: (a) a conceptual shift from safety as exposure reduction towards safety as enablement; and (b) renewed attention to evaluating the trauma-informedness of design processes, in tandem with the outcomes of designed objects on users' well-being., Comment: To appear at ACM CSCW 2025

Causality plays an important role in understanding intelligent behavior, and there is a wealth of literature on mathematical models for causality, most of which is focused on causal graphs. Causal graphs are a powerful tool for a wide range of applications, in particular when the relevant variables are known and at the same level of abstraction. However, the given variables can also be unstructured data, like pixels of an image. Meanwhile, the causal variables, such as the positions of objects in the image, can be arbitrary deterministic functions of the given variables. Moreover, the causal variables may form a hierarchy of abstractions, in which the macro-level variables are deterministic functions of the micro-level variables. Causal graphs are limited when it comes to modeling this kind of situation. In the presence of deterministic relationships there is generally no causal graph that satisfies both the Markov condition and the faithfulness condition. We introduce factored space models as an alternative to causal graphs which naturally represent both probabilistic and deterministic relationships at all levels of abstraction. Moreover, we introduce structural independence and establish that it is equivalent to statistical independence in every distribution that factorizes over the factored space. This theorem generalizes the classical soundness and completeness theorem for d-separation., Comment: 29 pages

Recent clinical results indicate that aberrant erythrocyte aggregation in hematological disorders is accompanied by endothelial damage and glycocalyx disruption, but the underlying biophysical mechanisms remain unclear. This study uses direct computational modeling to explore how red blood cell (RBC) aggregation impacts shear stress in small blood vessels, highlighting the increased risk of vascular damage. RBC aggregation creates a heterogeneous distribution, leading to variations in the cell-free layer thickness and fluctuating wall shear stress, especially near vessel walls. This effect aligns with experimental findings on endothelial disruption linked to RBC clustering near the wall, potentially reducing the protective glycocalyx layer. The power spectral density analysis of wall shear stress fluctuations reveals that, with RBC aggregation, there is a distinct peak near frequency f = 0.04, indicating increased fluctuations due to aggregated RBC clusters traveling close to the vessel wall. The presence of aberrant cells in blood disorders, modeled here by sickle cells, further amplifies these effects, as aggregation-enhanced margination drives sickle cells closer to vessel walls, exacerbating shear stress fluctuations and increasing the likelihood of vascular injury and inflammation. Simulations show that curved vascular geometry, with curvature accentuating RBC clustering near vessel walls, intensifies aggregation-induced wall shear stress fluctuations and increases the risk of vascular damage, particularly in sickle cell disease where sickle cells marginate closer to the wall.

We present a randomized algorithm for solving low-degree polynomial equation systems over finite fields faster than exhaustive search. In order to do so, we follow a line of work by Lokshtanov, Paturi, Tamaki, Williams, and Yu (SODA 2017), Bj\"orklund, Kaski, and Williams (ICALP 2019), and Dinur (SODA 2021). In particular, we generalize Dinur's algorithm for $\mathbb{F}_2$ to all finite fields, in particular the "symbolic interpolation" of Bj\"orklund, Kaski, and Williams, and we use an efficient trimmed multipoint evaluation and interpolation procedure for multivariate polynomials over finite fields by Van der Hoeven and Schost (AAECC 2013). The running time of our algorithm matches that of Dinur's algorithm for $\mathbb{F}_2$ and is significantly faster than the one of Lokshtanov et al. for $q>2$. We complement our results with tight conditional lower bounds that, surprisingly, we were not able to find in the literature. In particular, under the strong exponential time hypothesis, we prove that it is impossible to solve $n$-variate low-degree polynomial equation systems over $\mathbb{F}_q$ in time $O((q-\varepsilon)^{n})$. As a bonus, we show that under the counting version of the strong exponential time hypothesis, it is impossible to compute the number of roots of a single $n$-variate low-degree polynomial over $\mathbb{F}_q$ in time ${O((q-\varepsilon)^{n})}$; this generalizes a result of Williams (SOSA 2018) from $\mathbb{F}_2$ to all finite fields.

Preclinical diffusion MRI (dMRI) has proven value in methods development and validation, characterizing the biological basis of diffusion phenomena, and comparative anatomy. While dMRI enables in vivo non-invasive characterization of tissue, ex vivo dMRI is increasingly being used to probe tissue microstructure and brain connectivity. Ex vivo dMRI has several experimental advantages that facilitate high spatial resolution and high signal-to-noise ratio (SNR) images, cutting-edge diffusion contrasts, and direct comparison with histological data as a methodological validation. However, there are a number of considerations that must be made when performing ex vivo experiments. The steps from tissue preparation, image acquisition and processing, and interpretation of results are complex, with many decisions that not only differ dramatically from in vivo imaging of small animals, but ultimately affect what questions can be answered using the data. This work concludes a 3-part series of recommendations and considerations for preclinical dMRI. Herein, we describe best practices for dMRI of ex vivo tissue, with a focus on image pre-processing, data processing and model fitting, and tractography. In each section, we attempt to provide guidelines and recommendations, but also highlight areas for which no guidelines exist (and why), and where future work should lie. We end by providing guidelines on code sharing and data sharing, and point towards open-source software and databases specific to small animal and ex vivo imaging., Comment: Part 3 of 3 in "Considerations and recommendations for preclinical diffusion MRI"

The 1st Workshop on Data Contamination (CONDA 2024) focuses on all relevant aspects of data contamination in natural language processing, where data contamination is understood as situations where evaluation data is included in pre-training corpora used to train large scale models, compromising evaluation results. The workshop fostered a shared task to collect evidence on data contamination in current available datasets and models. The goal of the shared task and associated database is to assist the community in understanding the extent of the problem and to assist researchers in avoiding reporting evaluation results on known contaminated resources. The shared task provides a structured, centralized public database for the collection of contamination evidence, open to contributions from the community via GitHub pool requests. This first compilation paper is based on 566 reported entries over 91 contaminated sources from a total of 23 contributors. The details of the individual contamination events are available in the platform. The platform continues to be online, open to contributions from the community., Comment: https://huggingface.co/spaces/CONDA-Workshop/Data-Contamination-Database

Graph representation learning is central for the application of machine learning (ML) models to complex graphs, such as social networks. Ensuring `fair' representations is essential, due to the societal implications and the use of sensitive personal data. In this paper, we demonstrate how the parametrization of the \emph{CrossWalk} algorithm influences the ability to infer a sensitive attributes from node embeddings. By fine-tuning hyperparameters, we show that it is possible to either significantly enhance or obscure the detectability of these attributes. This functionality offers a valuable tool for improving the fairness of ML systems utilizing graph embeddings, making them adaptable to different fairness paradigms., Comment: Poster at ICML 2024 Workshop on the Next Generation of AI Safety

Deep learning provides powerful methods to impute structured information from large-scale, unstructured text and image datasets. For example, economists might wish to detect the presence of economic activity in satellite images, or to measure the topics or entities mentioned in social media, the congressional record, or firm filings. This review introduces deep neural networks, covering methods such as classifiers, regression models, generative AI, and embedding models. Applications include classification, document digitization, record linkage, and methods for data exploration in massive scale text and image corpora. When suitable methods are used, deep learning models can be cheap to tune and can scale affordably to problems involving millions or billions of data points.. The review is accompanied by a companion website, EconDL, with user-friendly demo notebooks, software resources, and a knowledge base that provides technical details and additional applications.

Social scientists and the general public often analyze contemporary events by drawing parallels with the past, a process complicated by the vast, noisy, and unstructured nature of historical texts. For example, hundreds of millions of page scans from historical newspapers have been noisily transcribed. Traditional sparse methods for searching for relevant material in these vast corpora, e.g., with keywords, can be brittle given complex vocabularies and OCR noise. This study introduces News Deja Vu, a novel semantic search tool that leverages transformer large language models and a bi-encoder approach to identify historical news articles that are most similar to modern news queries. News Deja Vu first recognizes and masks entities, in order to focus on broader parallels rather than the specific named entities being discussed. Then, a contrastively trained, lightweight bi-encoder retrieves historical articles that are most similar semantically to a modern query, illustrating how phenomena that might seem unique to the present have varied historical precedents. Aimed at social scientists, the user-friendly News Deja Vu package is designed to be accessible for those who lack extensive familiarity with deep learning. It works with large text datasets, and we show how it can be deployed to a massive scale corpus of historical, open-source news articles. While human expertise remains important for drawing deeper insights, News Deja Vu provides a powerful tool for exploring parallels in how people have perceived past and present.

Massive-scale historical document collections are crucial for social science research. Despite increasing digitization, these documents typically lack unique cross-document identifiers for individuals mentioned within the texts, as well as individual identifiers from external knowledgebases like Wikipedia/Wikidata. Existing entity disambiguation methods often fall short in accuracy for historical documents, which are replete with individuals not remembered in contemporary knowledgebases. This study makes three key contributions to improve cross-document coreference resolution and disambiguation in historical texts: a massive-scale training dataset replete with hard negatives - that sources over 190 million entity pairs from Wikipedia contexts and disambiguation pages - high-quality evaluation data from hand-labeled historical newswire articles, and trained models evaluated on this historical benchmark. We contrastively train bi-encoder models for coreferencing and disambiguating individuals in historical texts, achieving accurate, scalable performance that identifies out-of-knowledgebase individuals. Our approach significantly surpasses other entity disambiguation models on our historical newswire benchmark. Our models also demonstrate competitive performance on modern entity disambiguation benchmarks, particularly certain news disambiguation datasets.

In the U.S. historically, local newspapers drew their content largely from newswires like the Associated Press. Historians argue that newswires played a pivotal role in creating a national identity and shared understanding of the world, but there is no comprehensive archive of the content sent over newswires. We reconstruct such an archive by applying a customized deep learning pipeline to hundreds of terabytes of raw image scans from thousands of local newspapers. The resulting dataset contains 2.7 million unique public domain U.S. newswire articles, written between 1878 and 1977. Locations in these articles are georeferenced, topics are tagged using customized neural topic classification, named entities are recognized, and individuals are disambiguated to Wikipedia using a novel entity disambiguation model. To construct the Newswire dataset, we first recognize newspaper layouts and transcribe around 138 millions structured article texts from raw image scans. We then use a customized neural bi-encoder model to de-duplicate reproduced articles, in the presence of considerable abridgement and noise, quantifying how widely each article was reproduced. A text classifier is used to ensure that we only include newswire articles, which historically are in the public domain. The structured data that accompany the texts provide rich information about the who (disambiguated individuals), what (topics), and where (georeferencing) of the news that millions of Americans read over the course of a century. We also include Library of Congress metadata information about the newspapers that ran the articles on their front pages. The Newswire dataset is useful both for large language modeling - expanding training data beyond what is available from modern web texts - and for studying a diversity of questions in computational linguistics, social science, and the digital humanities., Comment: arXiv admin note: text overlap with arXiv:2306.17810, arXiv:2308.12477

In today's interconnected world, achieving reliable out-of-distribution (OOD) detection poses a significant challenge for machine learning models. While numerous studies have introduced improved approaches for multi-class OOD detection tasks, the investigation into multi-label OOD detection tasks has been notably limited. We introduce Spectral Normalized Joint Energy (SNoJoE), a method that consolidates label-specific information across multiple labels through the theoretically justified concept of an energy-based function. Throughout the training process, we employ spectral normalization to manage the model's feature space, thereby enhancing model efficacy and generalization, in addition to bolstering robustness. Our findings indicate that the application of spectral normalization to joint energy scores notably amplifies the model's capability for OOD detection. We perform OOD detection experiments utilizing PASCAL-VOC as the in-distribution dataset and ImageNet-22K or Texture as the out-of-distribution datasets. Our experimental results reveal that, in comparison to prior top performances, SNoJoE achieves 11% and 54% relative reductions in FPR95 on the respective OOD datasets, thereby defining the new state of the art in this field of study.

Clone-censor-weighting (CCW) is an analytic method for studying treatment regimens that are indistinguishable from one another at baseline without relying on landmark dates or creating immortal person time. One particularly interesting CCW application is estimating outcomes when starting treatment within specific time windows in observational data (e.g., starting a treatment within 30 days of hospitalization). In such cases, CCW estimates something fairly complex. We show how using CCW to study a regimen such as "start treatment prior to day 30" estimates the potential outcome of a hypothetical intervention where A) prior to day 30, everyone follows the treatment start distribution of the study population and B) everyone who has not initiated by day 30 initiates on day 30. As a result, the distribution of treatment initiation timings provides essential context for the results of CCW studies. We also show that if the exposure effect varies over time, ignoring exposure history when estimating inverse probability of censoring weights (IPCW) estimates the risk under an impossible intervention and can create selection bias. Finally, we examine some simplifying assumptions that can make this complex treatment effect more interpretable and allow everyone to contribute to IPCW.

AI-driven medical predictions with trustworthy confidence are essential for ensuring the responsible use of AI in healthcare applications. The growing capabilities of AI raise questions about their trustworthiness in healthcare, particularly due to opaque decision-making and limited data availability. This paper proposes a novel approach to address these challenges, introducing a Bayesian Monte Carlo Dropout model with kernel modelling. Our model is designed to enhance reliability on small medical datasets, a crucial barrier to the wider adoption of AI in healthcare. This model leverages existing language models for improved effectiveness and seamlessly integrates with current workflows. Extensive evaluations of public medical datasets showcase our model's superior performance across diverse tasks. We demonstrate significant improvements in reliability, even with limited data, offering a promising step towards building trust in AI-driven medical predictions and unlocking its potential to improve patient care.

Predicting legal judgments with reliable confidence is paramount for responsible legal AI applications. While transformer-based deep neural networks (DNNs) like BERT have demonstrated promise in legal tasks, accurately assessing their prediction confidence remains crucial. We present a novel Bayesian approach called BayesJudge that harnesses the synergy between deep learning and deep Gaussian Processes to quantify uncertainty through Bayesian kernel Monte Carlo dropout. Our method leverages informative priors and flexible data modelling via kernels, surpassing existing methods in both predictive accuracy and confidence estimation as indicated through brier score. Extensive evaluations of public legal datasets showcase our model's superior performance across diverse tasks. We also introduce an optimal solution to automate the scrutiny of unreliable predictions, resulting in a significant increase in the accuracy of the model's predictions by up to 27\%. By empowering judges and legal professionals with more reliable information, our work paves the way for trustworthy and transparent legal AI applications that facilitate informed decisions grounded in both knowledge and quantified uncertainty.

This paper presents Saharaline, an intervention designed to provide collective social support for teachers in low-income schools. Implemented as a WhatsApp-based helpline, Saharaline enables teachers to reach out for personalized, long-term assistance with a wide range of problems and stressors, including pedagogical, emotional, and technological challenges. Depending on the support needed, teachers' requests are routed to appropriate domain experts -- staff employed by educational non-profit organizations who understand teachers' on-the-ground realities -- who offer localized and contextualized assistance. Via a three-month exploratory deployment with 28 teachers in India, we show how Saharaline's design enabled a collective of diverse education experts to craft and deliver localized solutions that teachers could incorporate into their practice. We conclude by reflecting on the efficacy of our intervention in low-resource work contexts and provide recommendations to enhance collective social support interventions similar to Saharaline.

The emission of the pre-equilibrium particles during nuclear collisions at moderate beam energies is still an open question. This influences the properties of the compound nucleus but also changes the interpretation of the quasi-fission process. A systematic analysis of the data obtained by the FAZIA collaboration during a recent experiment with a neutron rich projectile is presented. The full range of charged particles detected in the experiment is within the limit of isotopic resolution of the FAZIA detector. Quasi-projectile (QP) fragments were detected in majority thanks to the forward angular acceptance of the experimental setup which was confirmed by introducing cuts based on the HIPSE event generator calculations. The main goal was to compare the experimental results with the HIPSE simulations after introducing these cuts to investigate the influence of the n-rich entrance channel on the QP fragment properties. More specifically, the lowering of N/Z of QP fragments with beam energy was found to be present since the initial phase of the reaction. Thus, pre-equilibrium emissions might be a possible candidate to explain such an effect., Comment: 10 pages, 10 figures

Plasmonic lattice nanostructures are of technological interest because of their capacity to manipulate light below the diffraction limit. Here, we present a detailed study of dark and bright modes in the visible and near-infrared energy regime of an inverted plasmonic honeycomb lattice by a combination of Au+ focused ion beam lithography with nanometric resolution, optical and electron spectroscopy, and finite-difference time-domain simulations. The lattice consists of slits carved in a gold thin film, exhibiting hotspots and a set of bright and dark modes. We proposed that some of the dark modes detected by electron energy-loss spectroscopy are caused by antiferroelectric arrangements of the slit polarizations with two times the size of the hexagonal unit cell. The plasmonic resonances take place within the 0.5_2 eV energy range, indicating that they could be suitable for a synergistic coupling with excitons in two-dimensional transition metal dichalcogenides materials or for designing nanoscale sensing platforms based on near-field enhancement over a metallic surface.

Given a triangulated category $D$ with an action of a fusion category $C$, we study the moduli space $Stab_{C}(D)$ of fusion-equivariant Bridgeland stability conditions on $D$. The main theorem is that the fusion-equivariant stability conditions form a closed, complex submanifold of the moduli space of stability conditions on $D$. As an application of this framework, we generalise a result of Macr\`{i}--Mehrotra--Stellari by establishing a homeomorphism between the space of $G$-invariant stability conditions on $D$ and the space of $rep(G)$-equivariant stability conditions on the equivariant category $D^G$. We also describe applications to the study of stability conditions associated to McKay quivers and to geometric stability conditions on free quotients of smooth projective varieties., Comment: v2: Updated the proof of lemma 3.17 (3.16 in previous version); new geometrical applications included; added an appendix dealing with the support property. v1: 30 pages. Comments more than welcomed!

Let $Y$ admit a rectangular Lefschetz decomposition of its derived category, and consider a cyclic cover $X\to Y$ ramified over a divisor $Z$. In a setting not considered by Kuznetsov and Perry, we define a subcategory $\mathcal{A}_Z$ of the equivariant derived category of $X$ which contains, rather than is contained in, $\mathrm{D}^{\mathrm{b}}(Z)$. We then show that the equivariant category of the Kuznetsov component of $X$ is decomposed into copies of $\mathcal{A}_Z$. As an application, we relate $\mathcal{A}_Z$ with the cohomology of $Z$ under some numerical assumptions. In particular, we obtain categorical Torelli theorems for the lowest degree prime Fano threefolds of index 1 and 2., Comment: 19 pages. Added a categorical Torelli theorem for prime Fano threefolds of index 2 and degree 1. Slightly weakened the generality assumptions. Other minor changes. Comments welcome!

This is the Proceedings of the 3rd International Workshop on Mining and Learning in the Legal Domain (MLLD-23) which took place in conjunction with the 32nd ACM International Conference on Information and Knowledge Management (CIKM-2023) at the University of Birmingham, Birmingham, UK on Sunday 22nd October 2023.

Billions of public domain documents remain trapped in hard copy or lack an accurate digitization. Modern natural language processing methods cannot be used to index, retrieve, and summarize their texts; conduct computational textual analyses; or extract information for statistical analyses, and these texts cannot be incorporated into language model training. Given the diversity and sheer quantity of public domain texts, liberating them at scale requires optical character recognition (OCR) that is accurate, extremely cheap to deploy, and sample-efficient to customize to novel collections, languages, and character sets. Existing OCR engines, largely designed for small-scale commercial applications in high resource languages, often fall short of these requirements. EffOCR (EfficientOCR), a novel open-source OCR package, meets both the computational and sample efficiency requirements for liberating texts at scale by abandoning the sequence-to-sequence architecture typically used for OCR, which takes representations from a learned vision model as inputs to a learned language model. Instead, EffOCR models OCR as a character or word-level image retrieval problem. EffOCR is cheap and sample efficient to train, as the model only needs to learn characters' visual appearance and not how they are used in sequence to form language. Models in the EffOCR model zoo can be deployed off-the-shelf with only a few lines of code. Importantly, EffOCR also allows for easy, sample efficient customization with a simple model training interface and minimal labeling requirements due to its sample efficiency. We illustrate the utility of EffOCR by cheaply and accurately digitizing 20 million historical U.S. newspaper scans, evaluating zero-shot performance on randomly selected documents from the U.S. National Archives, and accurately digitizing Japanese documents for which all other OCR solutions failed.

The College Completion Network conducted a systematic review of the research on advising strategies that support college completion. Several key findings can inform policy and practice: Holistic and comprehensive interventions that combine multiple components to meet student needs improve student outcomes. These types of interventions provide students with a comprehensive suite of advising, student services, academic support, and financial aid. Rigorous studies have consistently found long-term, positive effects on student outcomes. Proactive communication with sustained follow-up that spans a student's trajectory through college improves student outcomes. The review found that students benefit when colleges communicate proactively, pair outreach with consistent follow-up, and provide incentives for students to respond. Informational interventions, such as those delivered via text messaging or email campaigns, should provide targeted and actionable information with clear next steps. Student success can be bolstered when colleges provide a robust set of nonacademic supports through comprehensive case management. Now more than ever, college students require a range of nonacademic supports to ensure their success in college. Colleges can partner with case management organizations to connect students with mental health supports, childcare, transportation assistance, and access to food and other basic needs. As new evidence merges, practitioners can learn about the most effective approaches to providing nonacademic supports. [For "Advising for College Success: A Systematic Review of the Evidence. Appendix C," see ED626639. For "Advising for College Success: Policies, Practices, and Programs That Support Students on the Path to College Completion," see ED626635. For "Advising for College Success: Next Steps for Policymakers, Practitioners, and Researchers," see ED626636.]

Linking information across sources is fundamental to a variety of analyses in social science, business, and government. While large language models (LLMs) offer enormous promise for improving record linkage in noisy datasets, in many domains approximate string matching packages in popular softwares such as R and Stata remain predominant. These packages have clean, simple interfaces and can be easily extended to a diversity of languages. Our open-source package LinkTransformer aims to extend the familiarity and ease-of-use of popular string matching methods to deep learning. It is a general purpose package for record linkage with transformer LLMs that treats record linkage as a text retrieval problem. At its core is an off-the-shelf toolkit for applying transformer models to record linkage with four lines of code. LinkTransformer contains a rich repository of pre-trained transformer semantic similarity models for multiple languages and supports easy integration of any transformer language model from Hugging Face or OpenAI. It supports standard functionality such as blocking and linking on multiple noisy fields. LinkTransformer APIs also perform other common text data processing tasks, e.g., aggregation, noisy de-duplication, and translation-free cross-lingual linkage. Importantly, LinkTransformer also contains comprehensive tools for efficient model tuning, to facilitate different levels of customization when off-the-shelf models do not provide the required accuracy. Finally, to promote reusability, reproducibility, and extensibility, LinkTransformer makes it easy for users to contribute their custom-trained models to its model hub. By combining transformer language models with intuitive APIs that will be familiar to many users of popular string matching packages, LinkTransformer aims to democratize the benefits of LLMs among those who may be less familiar with deep learning frameworks.

Research involving at-risk users -- that is, users who are more likely to experience a digital attack or to be disproportionately affected when harm from such an attack occurs -- can pose significant safety challenges to both users and researchers. Nevertheless, pursuing research in computer security and privacy is crucial to understanding how to meet the digital-safety needs of at-risk users and to design safer technology for all. To standardize and bolster safer research involving such users, we offer an analysis of 196 academic works to elicit 14 research risks and 36 safety practices used by a growing community of researchers. We pair this inconsistent set of reported safety practices with oral histories from 12 domain experts to contribute scaffolded and consolidated pragmatic guidance that researchers can use to plan, execute, and share safer digital-safety research involving at-risk users. We conclude by suggesting areas for future research regarding the reporting, study, and funding of at-risk user research, Comment: 13 pages, 3 tables

Existing full text datasets of U.S. public domain newspapers do not recognize the often complex layouts of newspaper scans, and as a result the digitized content scrambles texts from articles, headlines, captions, advertisements, and other layout regions. OCR quality can also be low. This study develops a novel, deep learning pipeline for extracting full article texts from newspaper images and applies it to the nearly 20 million scans in Library of Congress's public domain Chronicling America collection. The pipeline includes layout detection, legibility classification, custom OCR, and association of article texts spanning multiple bounding boxes. To achieve high scalability, it is built with efficient architectures designed for mobile phones. The resulting American Stories dataset provides high quality data that could be used for pre-training a large language model to achieve better understanding of historical English and historical world knowledge. The dataset could also be added to the external database of a retrieval-augmented language model to make historical information - ranging from interpretations of political events to minutiae about the lives of people's ancestors - more widely accessible. Furthermore, structured article texts facilitate using transformer-based methods for popular social science applications like topic classification, detection of reproduced content, and news story clustering. Finally, American Stories provides a massive silver quality dataset for innovating multimodal layout analysis models and other multimodal applications.

We study slope-stable vector bundles and Bridgeland stability conditions on varieties which are a quotient of a smooth projective variety by a finite abelian group $G$ acting freely. We show there is a one-to-one correspondence between $\widehat{G}$-invariant geometric stability conditions on the quotient and $G$-invariant geometric stability conditions on the cover. We apply our results to describe a connected component inside the stability manifolds of free abelian quotients when the cover has finite Albanese morphism. This applies to varieties with non-finite Albanese morphism which are free abelian quotients of varieties with finite Albanese morphism, such as Beauville-type and bielliptic surfaces. This gives a partial answer to a question raised by Lie Fu, Chunyi Li, and Xiaolei Zhao: If a variety $X$ has non-finite Albanese morphism, does there always exist a non-geometric stability condition on $X$? We also give counterexamples to a conjecture of Fu-Li-Zhao concerning the Le Potier function, which characterises Chern classes of slope-semistable sheaves. As a result of independent interest, we give a description of the set of geometric stability conditions on an arbitrary surface in terms of a refinement of the Le Potier function. This generalises a result of Fu-Li-Zhao from Picard rank one to arbitrary Picard rank., Comment: 34 pages, minor changes