Incomplete Representation Research Articles

Part-Prototype Models in Medical Imaging: Applications and Current Challenges

Recent developments in Artificial Intelligence have increasingly focused on explainability research. The potential of Explainable Artificial Intelligence (XAI) in producing trustworthy computer-aided diagnosis systems and its usage for knowledge discovery are gaining interest in the medical imaging (MI) community to support the diagnostic process and the discovery of image biomarkers. Most of the existing XAI applications in MI are focused on interpreting the predictions made using deep neural networks, typically including attribution techniques with saliency map approaches and other feature visualization methods. However, these are often criticized for providing incorrect and incomplete representations of the black-box models’ behaviour. This highlights the importance of proposing models intentionally designed to be self-explanatory. In particular, part-prototype (PP) models are interpretable-by-design computer vision (CV) models that base their decision process on learning and identifying representative prototypical parts from input images, and they are gaining increasing interest and results in MI applications. However, the medical field has unique characteristics that could benefit from more advanced implementations of these types of architectures. This narrative review summarizes existing PP networks, their application in MI analysis, and current challenges.

Unsupervised Clustering-based 3D Static Scene Construction Using LiDAR Channel and Azimuth Angle

Abstract. Cameras are typically used at road intersections to collect data to perform object detection but struggle in low-light and harsh weather. On the other hand, Light Detection and Ranging (LiDAR) is used as a key technology in 3D vision systems. It gives the 3D point cloud, which includes accurate depth information, but its resolution is cost-dependent, with higher resolutions being more expensive. Deep learning-based method requires large, labelled dataset which increases the cost, time and accuracy depending on the model trained on labelled dataset. To perform object detection in point cloud data is a challenging task due to the incomplete representations, data sparsity and unavailability of training data. To overcome this by using an unsupervised approach, it is important to identify the static scene and then detect the moving object. This work incorporated a novel approach to construct static scene using azimuth angle and laser channel information using an unsupervised clustering approach. This work incorporates two modules I) data collection using VLP-16 LiDAR, and II) static scene construction using DBSCAN (density-based spatial clustering and noise) clustering-based approach. Data is collected at a 4-legged intersection and pre-processed to extract aggregated distances corresponding to the unique pair of azimuth angle and laser channel. DBSCAN is used to perform clustering on the aggregated distances, based on the highest silhouette score and lowest intra distance between points in cluster, static points are identified, and static scene constructed. The qualitative evaluation of method demonstrates that the algorithm effectively and accurately filters out background points.

Iterative Time Series Imputation by Maintaining Dependency Consistency

Data imputation is crucial in the analysis of incomplete time series, such as forecasting and classification, which involves learning dependencies among the observed values to infer missing ones. As there are no ground truths for missing values, the challenge of time series imputation lies in preventing the model from overfitting to spurious correlations. In this paper, we believe that ensuring dependency consistency between observed and imputed values in a sequence is paramount for data imputation. Based on this idea, we propose a model called IR 2 -Net 1 , which combines an incomplete representation mechanism ( IRM ) with an iterative reconstruction framework ( IRF ) to establish a closed-loop learning-validation imputation paradigm. Firstly, IRM facilitates the representation of dependencies in incomplete sequences while preserving their distributions and semantics, effectively preventing the model from capturing spurious correlations. Secondly, IRF enables the model to reconstruct identical complete sequences separately based on imputed and observed values, ensuring that the dependencies of imputed values remain consistent with those of the observed ones. We conduct experiments on four datasets and compare IR 2 -Net with seven state-of-the-art imputation models. The experiment results show that IR 2 -Net outperforms all the baselines by 4.1%-23.4% in terms of accuracy. Moreover, IRF and IRM are two general modules that can be easily integrated into two existing models, significantly enhancing their performance by 18.3%-42.0%.

Multi-view graph convolutional networks based on multi-source information fusion for mechanical fault diagnosis

Data-driven intelligent diagnosis methods have demonstrated significant success within prognostics and health management systems due to their powerful feature representation and non-linear fitting capabilities. However, most existing methods employ simple fusion strategies and neglecting data structure considerations, which lead to incomplete feature representations for final diagnostic decisions. In this article, we address these limitations by proposing a multiview graph convolutional network [Formula: see text] framework based on multi-source information fusion for fault diagnosis of key components in mechanical equipment. This is achieved through joint research on sample graph and feature graph learning. First, convolution neural networks named the SharedNet module are leveraged to extract multi-source features, and then the deep features are concatenated in a fully connected layer. After that, we construct multi-view initial graphs to combine the correlations between samples and features to mine rich graph topology structure. Then, we redesign the graph learning strategy to further increase the robustness and generalizability of the proposed method. Finally, extensive experimental results on the multi-information datasets show that our proposed [Formula: see text] can not only outperform the state-of-the-art methods of comparison but also extract multi-view graph topology and feature representations for fault diagnosis.

Journey towards unlocking intangible heritage’s tourism potential: supply-side perceptions, involvement and dynamics

ABSTRACT This research addresses the challenges and opportunities in integrating Intangible Cultural Heritage (ICH) into Cultural Heritage Tourism (CHT) offerings. Despite the growing interest in ICH, comprehensive research on its effective incorporation into CHT is lacking. Qualitative research was used employing interviews with 13 representatives from four case study areas known for their significant ICH offerings in the Dutch-speaking regions of Belgium and the Netherlands. Three research questions were formulated and tested to understand stakeholder collaboration, knowledge levels, and the complexities surrounding the term ‘ICH.’ The findings reveal a deficiency in stakeholder cooperation and varying levels of knowledge among supply-side stakeholders. The term ‘ICH’ itself was found to be complex but manageable through education and ethical sensitivity. The study recommends the involvement of an impartial mediator to facilitate stakeholder collaboration and suggests that heritage communities should be actively engaged in the process. Limitations include focusing on specific cultural and linguistic groups and incomplete stakeholder representation. The research has practical implications for policy formulation and Destination Management Organisations (DMOs) and offers a nuanced understanding of the challenges and solutions for integrating ICH into CHT.

SCAE: Structural Contrastive Auto-Encoder for Incomplete Multi-View Representation Learning

Describing an object from multiple perspectives often leads to incomplete data representation. Consequently, learning consistent representations for missing data from multiple views has emerged as a key focus in the realm of Incomplete Multi-view Representation Learning (IMRL). In recent years, various strategies, such as subspace learning, matrix decomposition, and deep learning, have been harnessed to develop numerous IMRL methods. In this article, our primary research revolves around IMRL, with a particular emphasis on addressing two main challenges. Firstly, we delve into the effective integration of intra-view similarity and contextual structure into a unified framework. Secondly, we explore the effective facilitation of information exchange and fusion across multiple views. To tackle these issues, we propose a deep learning approach known as Structural Contrastive Auto-Encoder (SCAE) to solve the challenges of IMRL. SCAE comprises two major components: intra-view structural representation learning and inter-view contrastive representation learning. The former involves capturing intra-view similarity by minimizing the Dirichlet energy of the feature matrix, while also applying spatial dispersion regularization to capture intra-view contextual structure. The latter encourages maximizing the mutual information of inter-view representations, facilitating information exchange and fusion across views. Experimental results demonstrate the efficacy of our approach in significantly enhancing model accuracy and robustly addressing IMRL problems. The code is available at https://github.com/limengran98/SCAE .

Multiplexed spatial mapping of chromatin features, transcriptome, and proteins in tissues.

The phenotypic and functional states of a cell are modulated by a complex interactive molecular hierarchy of multiple omics layers, involving the genome, epigenome, transcriptome, proteome, and metabolome. Spatial omics approaches have enabled the capture of information from different molecular layers directly in the tissue context. However, current technologies are limited to map one to two modalities at the same time, providing an incomplete representation of cellular identity. Such data is inadequate to fully understand complex biological systems and their underlying regulatory mechanisms. Here we present spatial-Mux-seq, a multi-modal spatial technology that allows simultaneous profiling of five different modalities, including genome-wide profiles of two histone modifications and open chromatin, whole transcriptome, and a panel of proteins at tissue scale and cellular level in a spatially resolved manner. We applied this technology to generate multi-modal tissue maps in mouse embryos and mouse brains, which discriminated more cell types and states than unimodal data. We investigated the spatiotemporal relationship between histone modifications, chromatin accessibility, gene and protein expression in neuron differentiation revealing the relationship between tissue organization, function, and gene regulatory networks. We were able to identify a radial glia spatial niche and revealed spatially changing gradient of epigenetic signals in this region. Moreover, we revealed previously unappreciated involvement of repressive histone marks in the mouse hippocampus. Collectively, the spatial multi-omics approach heralds a new era for characterizing tissue and cellular heterogeneity that single modality studies alone could not reveal.

Contextual visual and motion salient fusion framework for action recognition in dark environments

Infrared (IR) human action recognition (AR) exhibits resilience against shifting illumination conditions, changes in appearance, and shadows. It has valuable applications in numerous areas of future sustainable and smart cities including robotics, intelligent systems, security, and transportation. However, current IR-based recognition approaches predominantly concentrate on spatial or local temporal information and often overlook the potential value of global temporal patterns. This oversight can lead to incomplete representations of body part movements and prevent accurate optimization of a network. Therefore, a contextual-motion coalescence network (CMCNet) is proposed that operates in a streamlined and end-to-end manner for robust action representation in darkness in a near-infrared (NIR) setting. Initially, data are preprocessed to separate foreground, normalized, and resized. The framework employs two parallel modules: the contextual visual features learning module (CVFLM) for local feature extraction, and the temporal optical flow learning module (TOFLM) for acquiring motion dynamics. These modules focus on action-relevant regions used shift window-based operations to ensure accurate interpretation of motion information. The coalescence block harmoniously integrates the contextual and motion features within a unified framework. Finally, the temporal decoder module discriminatively identifies the boundaries of the action sequence. This sequence of steps ensures the synergistic optimization of both CVFLM and TOFLM and thorough competent feature extraction for precise AR. Evaluations of CMCNet are carried out on publicly available datasets, InfAR and NTURGB-D, where superior performance is achieved. Our model yields the highest average precision of 89% and 85% on these datasets, respectively, representing an improvement of 2.25% (on InfAR) compared to conventional methods operating at spatial and optical flow levels which underscores its efficacy.

HCBiLSTM-WOA: hybrid convolutional bidirectional long short-term memory with water optimization algorithm for autism spectrum disorder

Autism Spectrum Disorder (ASD) is a type of brain developmental disability that cannot be completely treated, but its impact can be reduced through early interventions. Early identification of neurological disorders will better assist in preserving the subjects’ physical and mental health. Although numerous research works exist for detecting autism spectrum disorder, they are cumbersome and insufficient for dealing with real-time datasets. Therefore, to address these issues, this paper proposes an ASD detection mechanism using a novel Hybrid Convolutional Bidirectional Long Short-Term Memory based Water Optimization Algorithm (HCBiLSTM-WOA). The prediction efficiency of the proposed HCBiLSTM-WOA method is investigated using real-time ASD datasets containing both ASD and non-ASD data from toddlers, children, adolescents, and adults. The inconsistent and incomplete representations of the raw ASD dataset are modified using preprocessing procedures such as handling missing values, predicting outliers, data discretization, and data reduction. The preprocessed data obtained is then fed into the proposed HCBiLSTM-WOA classification model to effectively predict the non-ASD and ASD classes. The initially randomly initialized hyperparameters of the HCBiLSTM model are adjusted and tuned using the water optimization algorithm (WOA) to increase the prediction accuracy of ASD. After detecting non-ASD and ASD classes, the HCBiLSTM-WOA method further classifies the ASD cases into respective stages based on the autistic traits observed in toddlers, children, adolescents, and adults. Also, the ethical considerations that should be taken into account when campaign ASD risk communication are complex due to the data privacy and unpredictability surrounding ASD risk factors. The fusion of sophisticated deep learning techniques with an optimization algorithm presents a promising framework for ASD diagnosis. This innovative approach shows potential in effectively managing intricate ASD data, enhancing diagnostic precision, and improving result interpretation. Consequently, it offers clinicians a tool for early and precise detection, allowing for timely intervention in ASD cases. Moreover, the performance of the proposed HCBiLSTM-WOA method is evaluated using various performance indicators such as accuracy, kappa statistics, sensitivity, specificity, log loss, and Area Under the Receiver Operating Characteristics (AUROC). The simulation results reveal the superiority of the proposed HCBiLSTM-WOA method in detecting ASD compared to other existing methods. The proposed method achieves a higher ASD prediction accuracy of about 98.53% than the other methods being compared.

A fine-grained RDF graph model for fuzzy spatiotemporal data

The uncertainty and spatiotemporal dynamics of information necessitate the urgent modeling of fuzzy spatiotemporal knowledge across various applications, with the Resource Description Framework (RDF) serving as a widely recognized data representation model. Existing models suffer from incomplete semantic representation and poor robustness in modeling fuzzy spatiotemporal data, e.g., lack of fuzziness in spatiotemporal semantics; lack of altitude in spatial semantics. Meanwhile, the algebraic framework regarding the model has not been investigated. Thus, in this paper, we first propose a new fine-grained fuzzy spatiotemporal RDF model. This model can represent fine-grained uncertain spatiotemporal semantics that may be associated with any element of a spatiotemporal RDF. We further define its graph algebraic operations. Note that we demonstrate the use of the algebraic operations for fuzzy spatiotemporal RDF querying. Finally, we establish a set of transformation rules for SPARQL query syntax to algebraic operations in fuzzy spatiotemporal RDF. In addition, we used experiments to evaluate the validity and rationality of our model.

Superclass-aware visual feature disentangling for generalized zero-shot learning

Zero-shot learning (ZSL) aims to learn a model trained on seen samples with the ability to recognize samples from unseen classes, while generalized ZSL (GZSL) takes a step closer to realistic scenarios by recognizing both of seen and unseen samples. The existing methods rely on the semantic descriptions as the side-information and conduct tight alignment between the visual and semantic spaces. However, the tight modality alignment may result in incomplete representations, leading to the loss of originally detailed and discriminative information. In this paper, we propose a simple yet effective superclass-aware visual feature disentangling method termed as SupVFD for GZSL. We use the neighbor relations of the semantic descriptions to define superclass and with the guide of superclass, our method disentangles visual features into discriminative and transferable factors. To this end, the semantic descriptions are used as implicit supervision, which preserves the valuable detailed and discriminative information in the visual features. The extensive experiments in both ZSL and GZSL settings prove our method outperforms the state-of-the-art methods for image object classification as well as video action recognition. Code is available at our github: https://github.com/changniu54/SupVFD-Master.

Stacked dynamic target regularization enhanced autoencoder for soft sensor in industrial processes

AbstractStacked autoencoders (SAEs) have great potential in developing soft sensors due to their excellent feature extraction capabilities. However, the pre‐training stage of SAE is unsupervised and some important information related to target variables may be discarded. Meanwhile, as the depth of the network increases, reconstruction errors continue to accumulate, resulting in incomplete feature representations of the original input. In addition, the dynamic nature of the data affects the predictive results of the model. To address these issues, the stacked dynamic target regularization enhanced autoencoder (SDTR‐EAE) method is proposed, which adds the DTR and the original input information layer by layer to enhance the feature extraction. To adapt to the dynamic changes in data and extract target‐related features, entropy weight grey relational analysis (EW‐GRA) is used as the DTR term to constrain the weight matrix and suppress irrelevant features. To reduce the accumulation of information loss during the reconstruction, an information enhancement layer is introduced, where the original inputs and the information of the hidden layers of previous DTR‐EAE units are added to the follow‐up DTR‐EAE unit. Finally, in the regression process, the DTR term is used again to fully utilize depth features for quality prediction and prevent overfitting. Experimental verifications using the debutanizer column and thermal power plant are conducted to validate the effectiveness of the proposed modelling method.

Impact of Procedure Risk vs Frailty on Outcomes of Elderly Patients Undergoing Emergency General Surgery: Results of a National Analysis.

The direct association between procedure risk and outcomes in elderly patients who undergo emergency general surgery (EGS) has not been analyzed. Studies only highlight the importance of frailty. A comprehensive analysis of relevant risk factors and their association with outcomes in elderly patients who undergo EGS is lacking. We hypothesized that procedure risk has a stronger association with relevant outcomes in elderly patients who undergo EGS compared with frailty. Elderly patients (age >65 years) undergoing EGS operative procedures were identified in the NSQIP database (2018 to 2020) and stratified based on the presence of frailty calculated by the Modified 5-Item Frailty Index (mFI-5; mFI 0 nonfrail, mFI 1 to 2 frail, and mFI ≥3 severely frail) and based on procedure risk. Multivariable regression models and receiving operative curve analysis were used to determine risk factors associated with outcomes. A total of 59,633 elderly patients who underwent EGS were classified into nonfrail (17,496; 29.3%), frail (39,588; 66.4%), and severely frail (2,549; 4.3%). There were 25,157 patients in the low-risk procedure group and 34,476 in the high-risk group. Frailty and procedure risk were associated with increased mortality, complications, failure to rescue, and readmissions. Differences in outcomes were greater when patients were stratified according to procedure risk compared with frailty stratification alone. Procedure risk had a stronger association with relevant outcomes in elderly patients who underwent EGS compared with frailty. Assessing frailty in the population of elderly patients who undergo EGS without adjusting for the type of procedure or procedure risk ultimately presents an incomplete representation of how frailty impacts patient-related outcomes.

PMiSLocMF: predicting miRNA subcellular localizations by incorporating multi-source features of miRNAs.

The microRNAs (miRNAs) play crucial roles in several biological processes. It is essential for a deeper insight into their functions and mechanisms by detecting their subcellular localizations. The traditional methods for determining miRNAs subcellular localizations are expensive. The computational methods are alternative ways to quickly predict miRNAs subcellular localizations. Although several computational methods have been proposed in this regard, the incomplete representations of miRNAs in these methods left the room for improvement. In this study, a novel computational method for predicting miRNA subcellular localizations, named PMiSLocMF, was developed. As lots of miRNAs have multiple subcellular localizations, this method was a multi-label classifier. Several properties of miRNA, such as miRNA sequences, miRNA functional similarity, miRNA-disease, miRNA-drug, and miRNA-mRNA associations were adopted for generating informative miRNA features. To this end, powerful algorithms [node2vec and graph attention auto-encoder (GATE)] and one newly designed scheme were adopted to process above properties, producing five feature types. All features were poured into self-attention and fully connected layers to make predictions. The cross-validation results indicated the high performance of PMiSLocMF with accuracy higher than 0.83, average area under the receiver operating characteristic curve (AUC) and area under the precision-recall curve (AUPR) exceeding 0.90 and 0.77, respectively. Such performance was better than all previous methods based on the same dataset. Further tests proved that using all feature types can improve the performance of PMiSLocMF, and GATE and self-attention layer can help enhance the performance. Finally, we deeply analyzed the influence of miRNA associations with diseases, drugs, and mRNAs on PMiSLocMF. The dataset and codes are available at https://github.com/Gu20201017/PMiSLocMF.

Dialogue-Rewriting Model Based on Transformer Pointer Extraction

In the multi-turn dialogue scenario, users commonly encounter challenges with pronoun referents and information omission, leading to semantically incomplete representations. These issues contribute to textual incoherence, as unclear referents and missing components hinder the semantic understanding of the spoken representations of text by machines. Currently, scholars frequently resort to multiple rounds of dialogue rewriting to address the semantic challenges posed by the machine comprehension of semantically missing texts with pronoun referents and information omissions. However, existing dialogue-rewriting methods often suffer from low precision and high latency in handling such texts. To mitigate these shortcomings, this paper proposes a Transformer-based dialogue-rewriting model that utilizes pointer extraction. The method leverages a Transformer pre-training model to effectively extract the potential semantic features of text and extract the key information of text by a pointer address. By extracting keywords and appropriately replacing or inserting text, the model restores referents and missing information. The experimental findings on an open-source Chinese multi-turn dialogue-rewriting dataset demonstrate the effectiveness of the proposed method in improving both the accuracy and efficiency of rewriting compared with existing methods. Specifically, the ROUGR-1 value increased by 2.9%, while the time consumption decreased by 50% compared with the benchmark method.

Individual-Level Digital Determinants of Health and Technology Acceptance of Patient Portals: Cross-Sectional Assessment.

Digital determinants of health (DDoH), including access to technological tools and digital health literacy, function independently as barriers to health. Assessment for DDoH is not routine within most health care systems, although addressing DDoH could help mitigate differential health outcomes and the digital divide. This study aims to assess the role of individual-level DDoH factors on patient enrollment in and use of the patient portal. We developed a multimodal, cross-sectional survey and deployed it to 11,424 individuals based on their preferred mode and language documented within the electronic medical record. Based on the Technology Acceptance Model, enrollment in and intent to use the patient portal were the outcomes of interest. Perceived usefulness and ease of use were assessed to determine construct validity, and exploratory investigations included individual-level DDoH, including internet and device access, availability of technological support, medical complexity, individual relationship with the health care system, and digital health literacy. Counts (n) and proportions (%) were used to describe response categories, and adjusted and unadjusted odds ratios are reported. This study included 1850 respondents (11,424 invited, 16.2% response rate), who were mostly female (1048/1850, 56.6%) and White (1240/1850, 67%), with an average age of 63 years. In the validation of the Technology Acceptance Model, measures of perceived ease of use (ie, using the patient portal will require a lot of mental effort; the patient portal will be very easy to use) and perceived usefulness (ie, the usefulness of the patient portal to send and receive messages with providers, schedule appointments, and refill medications) were positively associated with both enrollment in and intent to use the patient portal. Within adjusted models, perceived ease of use and perceived usefulness constructs, in addition to constructs of digital health literacy, knowing what health resources are available on the internet (adjusted odds ratio [aOR] 3.5, 95% CI 1.8-6.6), portal ease of use (aOR 2.8, 95% CI 1.6-5), and portal usefulness (aOR 2.4, 95% CI 1.4-4.2) were significantly associated with patient portal enrollment. Other factors associated with patient portal enrollment and intent to use included being comfortable reading and speaking English, reported use of the internet to surf the web or to send or receive emails, home internet access, and access to technology devices (computer, tablet, smartphone, etc). Assessing for and addressing individual-level DDoH, including digital health literacy, access to digital tools and technologies, and support of the relational aspects between patients, social support systems, and health care providers, could help mitigate disparities in health. By focusing efforts to assess for and address individual-level DDoH, an opportunity exists to improve digitally driven health care delivery outcomes like access and structural outcomes like bias built within algorithms created with incomplete representation across communities.

Digital footprints as means of measuring loneliness experience and embeddedness in social networks for designing digital mental health interventions

Introduction & BackgroundDespite existing research evidence for the negative influence of loneliness on people’s wellbeing, most studies focus on the experiences of older adults and the student population. Moreover, research concerning loneliness and digital footprints uses demographic proxies, as opposed to a behavioural focus, thus providing an incomplete representation of the phenomenon’s influence on the general population. Objectives & ApproachThis project aims to use people’s digital social media data (shared by the participants from their Facebook, Twitter, or Reddit accounts) to address people’s experiences of loneliness in order to provide guidance for the design of interventions catering to the improvement of the wellbeing of individuals. Screening the participants for loneliness levels using the UCLA Loneliness Scale and looking how these experiences differ cross-sectionally (25-65-year-olds; minorities) will help understand the following: social network structures, as shaped by loneliness experience; the dynamics within one’s social network; and the linguistic content of the relationships. Using digital footprints for language modelling and thematic analysis of digital language data shared by the participants, in addition to social network analysis (mapped out based on the individuals’ digital interactions) will allow insight into digital wellbeing. A traditional approach will be utilised alongside digital data analysis to address the issue of limited social media data representativeness - relationships formed in non-digital settings, along with the associated loneliness experiences, will be included. In addition to sharing their digital footprints, the participants will be surveyed and interviewed about their everyday offline and digital experiences of loneliness; as well as their social network structures and dynamics. The interview and survey data will be analysed using thematic analysis of text data and predictive models of quantitative survey responses; in addition to social network analysis of the relationships listed during the interview. Predictions of loneliness outcomes in relation to people’s digital and offline behaviour; and correlations between loneliness experiences and social network dynamics will be made from the data. Relevance to Digital FootprintsThe project focuses on people’s experiences of loneliness in both digital and offline settings utilising the analysis of digital footprints from social media and traditional survey- and interview-based methodology. This approach will allow to gain insight into the similarities and differences between social network structures and dynamics as well as loneliness experiences in digital and real-world relationships as these inevitably interplay in everyday life. The inclusion of digital footprints data will allow to measure and predict loneliness impact on mental health and digital social behaviour; and design tailored digital wellbeing interventions in the future. Conclusions & ImplicationsThe results will serve as basis for designing innovative loneliness interventions for both the public and counsellors to offer tailored wellbeing support. Additionally, the project aims to promote loneliness awareness. Incorporating a variety of experiences from different social groups will enable inclusive, user-centred approach.

The Soleymani and Collins Obstetric morbidity score (SaCOMS): A quantitative tool for measuring maternal morbidity from complex obstetric surgery such as placenta accreta spectrum (PAS)

It is currently very difficult to compare different management strategies for complex obstetric surgery, such as hysterectomy for severe Placenta Accreta Spectrum (PAS), as there is no widely accepted consensus for the classification of maternal surgical morbidity. Many studies focus on the amount of blood products transfused or admission to intensive care units (ICU). However, these are dependent on local policies and available resources. It also gives an incomplete representation of the entire ‘patient journey’ after they leave the operating room. Subsequent repeat procedures for lower urinary track damage is arguably worse from the woman’s perspective than a short stay on an intensive care unit (ICU) for observation.We suggest a version of the Clavien-Dindo morbidity classification specific to obstetrics. Then employ it to build a quantitative morbidity score which aims to reflect the whole ‘patient experience’ including the post-operative pathway. We then demonstrate the utility of this system in a cohort of women with Placenta Accreta Spectrum (PAS).The Clavien-Dindo classification was modified to reflect obstetric procedures and a quantitative morbidity measure, the Soleymani and Collins Obstetric Morbidity Score (SaCOMS), was developed based on this. Both were then validated using a survey-based consultation of a panel of experts in PAS and retrospectively applied to a cohort of 54 women who underwent caesarean hysterectomy for PAS.Clinicians with expertise in PAS believe that the Modified Obstetric Clavien-Dindo classification system and the novel SaCOMS tool can improve assessment of maternal morbidity, and better reflect the ‘patient experience’. Application of the classification system to a single-centre PAS cohort suggested that surgery by gynecologic-oncology surgeons may be associated with decreased incidence and cumulative morbidity outcomes for women with PAS, especially those with the most severe presentation.This study presents a clinically useful obstetric-specific classification system for surgical morbidity. SaCOMS also provides a quantitative reflection of the full patient- journey experienced as a result of surgical complications enabling a more patient-centered representation of morbidity.

Gradient-based joint inversion of point-source moment tensor and station-specific time-shifts

SUMMARY The misalignment of the observation and predicted waveforms in regional moment tensor inversion is mainly due to seismic models’ incomplete representation of the Earth's heterogeneities. Current moment tensor inversion techniques, allowing station-specific time-shifts to account for the model error, are computationally expensive. Here, we propose a gradient-based method to jointly invert moment-tensor parameters, centroid depth and unknown station-specific time-shifts utilizing the modern functionalities in deep learning frameworks. A $L_2^2$ misfit function between predicted synthetic and time-shifted observed seismograms is defined in the spectral domain, which is differentiable to all unknowns. The inverse problem is solved by minimizing the misfit function with a gradient descent algorithm. The method's feasibility, robustness and scalability are demonstrated using synthetic experiments and real earthquake data in the Long Valley Caldera, California. This work presents an example of fresh opportunities to apply advanced computational infrastructures developed in deep learning to geophysical problems.

Just War and Judgment in Fratelli Tutti

For decades the papal tradition has renounced the term ‘war’ as something around which to build an ethical approach. One can sympathize with this: resort to war seems the consequence of ethical failure and brings in its train a host of brutalities including rape, torture, and murder that harm both victims and perpetrators. But that view of ‘war’ is an incomplete representation of the possibilities of the uses of force to secure legitimate political goods. Thus the popes have struggled to maintain a clear voice in the face of abject tragedies like the Russian invasion of Ukraine: on the one hand, Pope Francis condemns war in an almost absolute manner; on the other, he recognizes the legitimacy of Ukrainian defensive uses of force. In so doing, of course, he merely abides the so-called ‘just war theory’ he seems to have discarded. My contribution will focus on the current state of the just war in Catholic teaching. I will maintain the state remains where it has been since at least the middle of the twentieth century: skeptical about the capacity of states to judge in terms of international law, but cognizant of their right to defend themselves against ongoing aggression.