Search

Coronary artery calcium (CAC) is highly predictive of cardiovascular events. While millions of chest CT scans are performed annually in the United States, CAC is not routinely quantified from scans done for non-cardiac purposes. A deep learning algorithm was developed using 446 expert segmentations to automatically quantify CAC on non-contrast, non-gated CT scans (AI-CAC). Our study differs from prior works as we leverage imaging data across the Veterans Affairs national healthcare system, from 98 medical centers, capturing extensive heterogeneity in imaging protocols, scanners, and patients. AI-CAC performance on non-gated scans was compared against clinical standard ECG-gated CAC scoring. Non-gated AI-CAC differentiated zero vs. non-zero and less than 100 vs. 100 or greater Agatston scores with accuracies of 89.4% (F1 0.93) and 87.3% (F1 0.89), respectively, in 795 patients with paired gated scans within a year of a non-gated CT scan. Non-gated AI-CAC was predictive of 10-year all-cause mortality (CAC 0 vs. >400 group: 25.4% vs. 60.2%, Cox HR 3.49, p < 0.005), and composite first-time stroke, MI, or death (CAC 0 vs. >400 group: 33.5% vs. 63.8%, Cox HR 3.00, p < 0.005). In a screening dataset of 8,052 patients with low-dose lung cancer-screening CTs (LDCT), 3,091/8,052 (38.4%) individuals had AI-CAC >400. Four cardiologists qualitatively reviewed LDCT images from a random sample of >400 AI-CAC patients and verified that 527/531 (99.2%) would benefit from lipid-lowering therapy. To the best of our knowledge, this is the first non-gated CT CAC algorithm developed across a national healthcare system, on multiple imaging protocols, without filtering intra-cardiac hardware, and compared against a strong gated CT reference. We report superior performance relative to previous CAC algorithms evaluated against paired gated scans that included patients with intra-cardiac hardware.

We consider a general class of elliptic equations on hypercomplex manifolds which includes the quaternionic Monge-Amp\`ere equation, the quaternionic Hessian equation and the Monge-Amp\`ere equation for quaternionic $(n-1)$-plurisubharmonic functions. We prove that under suitable assumptions the solutions to these equations on hyperk\"ahler manifolds satisfy a $C^{2,\alpha}$ a priori estimate., Comment: 14 pages

Short-term earthquake clustering is one of the most important features of seismicity. Clusters are identified using various techniques, generally deterministic and based on spatio-temporal windowing. Conversely, the leading rail in short-term earthquake forecasting has a probabilistic view of clustering, usually based on the Epidemic Type Aftershock Sequence (ETAS) models. In this study we compare seismic clusters, identified by two different deterministic window-based techniques, with the ETAS probabilities associated with any event in the clusters, thus investigating the consistency between deterministic and probabilistic approaches. The comparison is performed by considering, for each event in an identified cluster, the corresponding probability of being independent and the expected number of triggered events according to ETAS. Results show no substantial differences between the cluster identification procedures, and an overall consistency between the identified clusters and the relative events' ETAS probabilities.

The advanced machine learning algorithm NESTORE (Next STrOng Related Earthquake) was developed to forecast strong aftershocks in earthquake sequences and has been successfully tested in Italy, western Slovenia, Greece, and California. NESTORE calculates the probability of aftershocks reaching or exceeding the magnitude of the main earthquake minus one and classifies clusters as type A or B based on a 0.5 probability threshold. In this study, NESTORE was applied to Japan using data from the Japan Meteorological Agency catalog (1973-2024). Due to Japan's high seismic activity and class imbalance, new algorithms were developed to complement NESTORE. The first is a hybrid cluster identification method using ETAS-based stochastic declustering and deterministic graph-based selection. The second, REPENESE (RElevant features, class imbalance PErcentage, NEighbour detection, SElection), is optimized for detecting outliers in skewed class distributions. A new seismicity feature was proposed, showing good results in forecasting cluster classes in Japan. Trained with data from 1973 to 2004 and tested from 2005 to 2023, the method correctly forecasted 75% of A clusters and 96% of B clusters, achieving a precision of 0.75 and an accuracy of 0.94 six hours after the mainshock. It accurately classified the 2011 T\=ohoku event cluster. Near-real-time forecasting was applied to the sequence after the April 17, 2024 M6.6 earthquake in Shikoku, classifying it as a "Type B cluster," with validation expected on October 31, 2024., Comment: submitted to Geoscience Frontiers

Parameter learning is a crucial task in the field of Statistical Relational Artificial Intelligence: given a probabilistic logic program and a set of observations in the form of interpretations, the goal is to learn the probabilities of the facts in the program such that the probabilities of the interpretations are maximized. In this paper, we propose two algorithms to solve such a task within the formalism of Probabilistic Answer Set Programming, both based on the extraction of symbolic equations representing the probabilities of the interpretations. The first solves the task using an off-the-shelf constrained optimization solver while the second is based on an implementation of the Expectation Maximization algorithm. Empirical results show that our proposals often outperform existing approaches based on projected answer set enumeration in terms of quality of the solution and in terms of execution time. The paper has been accepted at the ICLP2024 conference and is under consideration in Theory and Practice of Logic Programming (TPLP)., Comment: The paper has been accepted at the ICLP2024 conference and is under consideration in Theory and Practice of Logic Programming (TPLP)

Context: Software specifications are usually written in natural language and may suffer from imprecision, ambiguity, and other quality issues, called thereafter, requirement smells. Requirement smells can hinder the development of a project in many aspects, such as delays, reworks, and low customer satisfaction. From an industrial perspective, we want to focus our time and effort on identifying and preventing the requirement smells that are of high interest. Aim: This paper aims to characterise 12 requirements smells in terms of frequency, severity, and effects. Method: We interviewed ten experienced practitioners from different divisions of a large international company in the safety-critical domain called MBDA Italy Spa. Results: Our interview shows that the smell types perceived as most severe are Ambiguity and Verifiability, while as most frequent are Ambiguity and Complexity. We also provide a set of six lessons learnt about requirements smells, such as that effects of smells are expected to differ across smell types. Conclusions: Our results help to increase awareness about the importance of requirement smells. Our results pave the way for future empirical investigations, ranging from a survey confirming our findings to controlled experiments measuring the effect size of specific requirement smells.

We provide a detailed treatment of special hyperhermitian metrics on hypercomplex manifolds. The quaternionic Gauduchon and quaternionic balanced conditions are investigated at length: we describe their properties and characterize their existence. We also consider strong HKT metrics proving an incompatibility result with balanced hyperhermitian metrics. Additionally, we introduce an Einstein-type condition and determine basic properties and obstructions. Finally, we collect several examples and counterexamples regarding the existence of these types of special metrics as well as two constructions that allow to build new examples., Comment: 50 pages. Comments are welcome, minor changes

Curated datasets for healthcare are often limited due to the need of human annotations from experts. In this paper, we present MedEval, a multi-level, multi-task, and multi-domain medical benchmark to facilitate the development of language models for healthcare. MedEval is comprehensive and consists of data from several healthcare systems and spans 35 human body regions from 8 examination modalities. With 22,779 collected sentences and 21,228 reports, we provide expert annotations at multiple levels, offering a granular potential usage of the data and supporting a wide range of tasks. Moreover, we systematically evaluated 10 generic and domain-specific language models under zero-shot and finetuning settings, from domain-adapted baselines in healthcare to general-purposed state-of-the-art large language models (e.g., ChatGPT). Our evaluations reveal varying effectiveness of the two categories of language models across different tasks, from which we notice the importance of instruction tuning for few-shot usage of large language models. Our investigation paves the way toward benchmarking language models for healthcare and provides valuable insights into the strengths and limitations of adopting large language models in medical domains, informing their practical applications and future advancements., Comment: Accepted to EMNLP 2023. Camera-ready version: updated IRB, added more evaluation results on LLMs such as GPT4, LLaMa2, and LLaMa2-chat

Medical image classification is a critical problem for healthcare, with the potential to alleviate the workload of doctors and facilitate diagnoses of patients. However, two challenges arise when deploying deep learning models to real-world healthcare applications. First, neural models tend to learn spurious correlations instead of desired features, which could fall short when generalizing to new domains (e.g., patients with different ages). Second, these black-box models lack interpretability. When making diagnostic predictions, it is important to understand why a model makes a decision for trustworthy and safety considerations. In this paper, to address these two limitations, we propose a new paradigm to build robust and interpretable medical image classifiers with natural language concepts. Specifically, we first query clinical concepts from GPT-4, then transform latent image features into explicit concepts with a vision-language model. We systematically evaluate our method on eight medical image classification datasets to verify its effectiveness. On challenging datasets with strong confounding factors, our method can mitigate spurious correlations thus substantially outperform standard visual encoders and other baselines. Finally, we show how classification with a small number of concepts brings a level of interpretability for understanding model decisions through case studies in real medical data., Comment: 18 pages, 12 figures

This paper focuses on the problem of finding a continuous extension of the hypercomplex logarithm along a path. While a branch of the complex logarithm can be defined in a small open neighbourhood of a strictly negative real point, no continuous branch of the hypercomplex logarithm can be defined in any open set $A\subset \mathbb K\setminus \{0\}$ which contains a strictly negative real point $x_0$ (here $\mathbb K$ represents the algebra of quaternions or octonions). To overcome these difficulties, we introduced the logarithmic manifold $\mathscr E_\mathbb K^+$ and then showed that if $q\in\mathbb K,\ q=x+Iy$ then $E(x+Iy) %= (\exp (x + Iy), Iy) = (\exp x \cos y + I\exp x \sin y, Iy)$ is an immersion and a diffeomorphism between $\mathbb K$ and $\mathscr E_\mathbb K^+$. In this paper, we consider lifts of paths in $\mathbb K\setminus\{0\}$ to the logarithmic manifold $\mathscr{E}^+_\mathbb K$; even though $\mathbb K \setminus \{0\}$ is simply connected, in general, given a path in $\mathbb K \setminus \{0\}$, the existence of a lift of this path to $\mathscr{E}^+_\mathbb K$ is not guaranteed. There is an obvious equivalence between the problem of lifting a path in $\mathbb K \setminus \{0\}$ and the one of finding a continuation of the hypercomplex logarithm $\log_{\mathbb K}$ along this path., Comment: 30 pages, 4 figures

Large Language Models (LLMs) exploit fine-tuning as a technique to adapt to diverse goals, thanks to task-specific training data. Task specificity should go hand in hand with domain orientation, that is, the specialization of an LLM to accurately address the tasks of a given realm of interest. However, models are usually fine-tuned over publicly available data or, at most, over ground data from databases, ignoring business-level definitions and domain experience. On the other hand, Enterprise Knowledge Graphs (EKGs) are able to capture and augment such domain knowledge via ontological reasoning. With the goal of combining LLM flexibility with the domain orientation of EKGs, we propose a novel neurosymbolic architecture that leverages the power of ontological reasoning to build task- and domain-specific corpora for LLM fine-tuning., Comment: Accepted at RuleML 2023

Sharing medical reports is essential for patient-centered care. A recent line of work has focused on automatically generating reports with NLP methods. However, different audiences have different purposes when writing/reading medical reports -- for example, healthcare professionals care more about pathology, whereas patients are more concerned with the diagnosis ("Is there any abnormality?"). The expectation gap results in a common situation where patients find their medical reports to be ambiguous and therefore unsure about the next steps. In this work, we explore the audience expectation gap in healthcare and summarize common ambiguities that lead patients to be confused about their diagnosis into three categories: medical jargon, contradictory findings, and misleading grammatical errors. Based on our analysis, we define a disambiguation rewriting task to regenerate an input to be unambiguous while preserving information about the original content. We further propose a rewriting algorithm based on contrastive pretraining and perturbation-based rewriting. In addition, we create two datasets, OpenI-Annotated based on chest reports and VA-Annotated based on general medical reports, with available binary labels for ambiguity and abnormality presence annotated by radiology specialists. Experimental results on these datasets show that our proposed algorithm effectively rewrites input sentences in a less ambiguous way with high content fidelity. Our code and annotated data are released to facilitate future research., Comment: Accepted to AAAI 2023. 13 pages including 4-page supplementary materials

We show that the parabolic quaternionic Monge-Amp\`ere equation on a compact hyperk\"ahler manifold has always a long-time solution which once normalized converges smoothly to a solution of the quaternionic Monge-Amp\`ere equation. This is the same setting in which Dinew and Sroka prove the conjecture of Alesker and Verbitsky. We also introduce an analogue of the Chern-Ricci flow in hyperhermitian manifolds., Comment: This is a new version of the paper "The Calabi-Yau theorem on Hypercomplex manifolds", which was withdrawn due to a crucial mistake in a key Lemma. In this new version we fix the mistake by introducing the extra assumption of the existence of a background hyperk\"ahler metric

We present a new method for the analysis of smoothly varying tapers, transitions and filters in rectangular waveguides. With this aim, we apply a Hierarchical Model (HiMod) reduction to the vector Helmholtz equation. We exploit a suitable coordinate transformation and, successively, we use the waveguide modes as a basis for the HiMod expansion. We show that accurate results can be obtained with an impressive speed-up factor when compared with standard commercial codes based on a three-dimensional finite element discretization.