Explicit Domain Research Articles

Knowledge Aware Automated Health Claims Processing with Medical Ontologies and Large Language Models

One of the major challenges in automating health insurance claims processing lies in the complexity involved in validating an incoming claim's medical diagnoses against its policy Underwriting (UW) exclusions. Termed UW Exclusion Detection, this process ensures claims are only paid out if their diagnoses are not medically associated with conditions excluded under the policy. Medical diagnoses in health insurance claims are typically represented by the International Classification of Disease (ICD) codes, established by the World Health Organization. For example, given a policy that excludes "all respiratory illness". A claim with the ICD code J45 (Asthma) will be subject to rejection as J45 is a respiratory-related diagnosis that falls within the scope of the policy's UW exclusion. The key challenge in automating this process lies in the wide range of available ICD codes. The ICD-10-CM coding scheme consists of over 40,000 codes, which often results in scenarios where codes encountered during inference are absent from the training data. These unseen ICD codes limit the effectiveness of data-driven approaches, which depend on the training data to discern medically relevant associations between UW exclusions and ICD codes. This underscores the need to supplement data-driven approaches with additional domain knowledge. We hypothesize that integrating implicit medical domain knowledge inherent in Large Language Models (LLMs) with explicit domain knowledge from medical ontologies, will enhance data-driven approaches for UW Exclusion Detection. Thoroughly validated on real-world health insurance claims data, our proposed approach proved effective in accurately establishing medically relevant associations between UW exclusions and ICD codes.

Integrating Domain Knowledge in Multi-Source Classification Tasks

This work presents an extended investigation into multi-domain learning techniques within the context of image and audio classification, with a focus on the latter. In machine learning, collections of data obtained or generated under similar conditions are referred to as domains or data sources. However, the distinct acquisition or generation conditions of these data sources are often overlooked, despite their potential to significantly impact model generalization. Multi-domain learning addresses this challenge by seeking effective methods to train models to perform adequately across all domains seen during the training process. Our study explores a range of model-agnostic multi-domain learning techniques that leverage explicit domain information alongside class labels. Specifically, we delve into three distinct methodologies: a general approach termed Stew, which involves mixing all available data indiscriminately; and two batch domain-regularization methods: Balanced Domains and Loss Sum. These methods are evaluated through several experiments conducted on datasets featuring multiple data sources for audio and image classification tasks. Our findings underscore the importance of considering domain-specific information during the training process. We demonstrate that the application of the Loss Sum method yields notable improvements in model performance (0.79 F1-Score) compared to conventional approaches that blend data from all available domains (0.62 F1-Score). By examining the impact of different multi-domain learning techniques on classification tasks, this study contributes to a deeper understanding of effective strategies for leveraging domain knowledge in machine learning model training.

Structure enhanced prototypical alignment for unsupervised cross-domain node classification

Graph Neural Networks (GNNs) have demonstrated remarkable success in graph node classification task. However, their performance heavily relies on the availability of high-quality labeled data, which can be time-consuming and labor-intensive to acquire for graph-structured data. Therefore, the task of transferring knowledge from a label-rich graph (source domain) to a completely unlabeled graph (target domain) becomes crucial. In this paper, we propose a novel unsupervised graph domain adaptation framework called Structure Enhanced Prototypical Alignment (SEPA), which aims to learn domain-invariant representations on non-IID (non-independent and identically distributed) data. Specifically, SEPA captures class-wise semantics by constructing a prototype-based graph and introduces an explicit domain discrepancy metric to align the source and target domains. The proposed SEPA framework is optimized in an end-to-end manner, which could be incorporated into various GNN architectures. Experimental results on several real-world datasets demonstrate that our proposed framework outperforms recent state-of-the-art baselines with different gains.

Device-edge collaborative occluded face recognition method based on cross-domain feature fusion

The lack of facial features caused by wearing masks degrades the performance of facial recognition systems. Traditional occluded face recognition methods cannot integrate the computational resources of the edge layer and the device layer. Besides, previous research fails to consider the facial characteristics including occluded and unoccluded parts. To solve the above problems, we put forward a device-edge collaborative occluded face recognition method based on cross-domain feature fusion. Specifically, the device-edge collaborative face recognition architecture gets the utmost out of maximizes device and edge resources for real-time occluded face recognition. Then, a cross-domain facial feature fusion method is presented which combines both the explicit domain and the implicit domain facial. Furthermore, a delay-optimized edge recognition task scheduling method is developed that comprehensively considers the task load, computational power, bandwidth, and delay tolerance constraints of the edge. This method can dynamically schedule face recognition tasks and minimize recognition delay while ensuring recognition accuracy. The experimental results show that the proposed method achieves an average gain of about 21% in recognition latency, while the accuracy of the face recognition task is basically the same compared to the baseline method.

KAMEL: Knowledge Aware Medical Entity Linkage to Automate Health Insurance Claims Processing

Automating the processing of health insurance claims to achieve "Straight-Through Processing" is one of the holy grails that all insurance companies aim to achieve. One of the major impediments to this automation is the difficulty in establishing the relationship between the underwriting exclusions that a policy has and the incoming claim's diagnosis information. Typically, policy underwriting exclusions are captured in free-text such as "Respiratory illnesses are excluded due to a pre-existing asthma condition". A medical claim coming from a hospital would have the diagnosis represented using the International Classification of Disease (ICD) codes from the World Health Organization. The complex and labour-intensive task of establishing the relationship between free-text underwriting exclusions in health insurance policies and medical diagnosis codes from health insurance claims is critical towards determining if a claim should be rejected due to underwriting exclusions. In this work, we present a novel framework that leverages both explicit and implicit domain knowledge present in medical ontologies and pre-trained language models respectively, to effectively establish the relationship between free-text describing medical conditions present in underwriting exclusions and the ICD-10CM diagnosis codes in health insurance claims. Termed KAMEL (Knowledge Aware Medical Entity Linkage), our proposed framework addresses the limitations faced by prior approaches when evaluated on real-world health insurance claims data. Our proposed framework have been deployed in several multi-national health insurance providers to automate their health insurance claims.

Region-Based Unsupervised Low-Light Image Enhancement in the Wild With Explicit Domain Supervision

Region-Based Unsupervised Low-Light Image Enhancement in the Wild With Explicit Domain Supervision

A state-based inverse reinforcement learning approach to model activity-travel choices behavior with reward function recovery

Behaviorally oriented activity-travel choices (ATC) modeling is a principal part of travel demand analysis. Traditional econometric and rule-based methods require explicit model structures and complex domain knowledge. While several recent studies used machine learning models, especially adversarial inverse reinforcement learning (IRL) models, to learn potential ATC patterns with less expert-designed settings, they lack a clear representation of rational ATC behavior. In this study, we propose a data-driven IRL framework based on the maximum causal approach to minimize f-divergences between expert and agent state marginal distributions, which provides a more sample-efficient measurement. In addition, we specify a separate state-only reward function and derive an analytical gradient of the f-divergence objective with respect to reward parameters to ensure good convergences. The method can recover a stationary reward function, which assures the agent to get close to the expert behavior when training from scratch. We validate the proposed model using cellular signaling data from Chongqing, China by comparing with baseline models (behavior cloning, policy-based, and reward-based models) in aspects of policy performance comparison, reward recovery, and reward transfer tasks. The experiment results indicate that the proposed model outperforms existing methods and is relatively less sensitive to the number of expert demonstrations. Qualitative analyses are provided on the fundamental ATC preferences on different features given the reward function recovered from the observed mobility trajectories, and on the learning behaviors under different choices of f-divergence.

Improving deep learning with prior knowledge and cognitive models: A survey on enhancing explainability, adversarial robustness and zero-shot learning

We review current and emerging knowledge-informed and brain-inspired cognitive systems for realizing adversarial defenses, eXplainable Artificial Intelligence (XAI), and zero-shot or few-shot learning. Data-driven machine learning models have achieved remarkable performance and demonstrated capabilities surpassing humans in many applications. Yet, their inability to exploit domain knowledge leads to serious performance limitations in practical applications. In particular, deep learning systems are exposed to adversarial attacks, which can trick them into making glaringly incorrect decisions. Moreover, complex data-driven models typically lack interpretability or explainability, i.e., their decisions cannot be understood by human subjects. Furthermore, models are usually trained on standard datasets with a closed-world assumption. Hence, they struggle to generalize to unseen cases during inference in practical open-world environments, thus, raising the zero- or few-shot generalization problem. Although many conventional solutions exist, explicit domain knowledge, brain-inspired neural networks and cognitive architectures offer powerful new dimensions towards alleviating these problems. Prior knowledge is represented in appropriate forms like mathematical relations, logic rules, knowledge graphs, and large language models (LLMs). and incorporated in deep learning frameworks to improve performance. Brain-inspired cognition methods use computational models that mimic the human brain to enhance intelligent behavior in artificial agents and autonomous robots. Ultimately, these models achieve better explainability, higher adversarial robustness and data-efficient learning, and can, in turn, provide insights for cognitive science and neuroscience—that is, to deepen human understanding on how the brain works in general, and how it handles these problems.

LegalHTML: Semantic mark-up of legal acts using web technologies

We introduce here LegalHTML, an extension of the HTML language thought for representing legal acts. LegalHTML has been conceived in the context of an exploratory study conducted for the Publications Office of the European Union, with the objective of overcoming the proliferation of formats for the electronic redaction of legal acts, dedicated to different steps of the editorial process (e.g. first draft, content editing, proof reading, introducing semantics, publishing) and of realizing a model and a language that could bind all processes and exigencies under a common umbrella. LegalHTML satisfies these requirements by providing an explicit domain language addressing all structural aspects of an act, such as articles, paragraphs, items, references and an associated ontology (foreseeing both inline annotations through RDFa and explicit RDF code within script elements) providing rich semantics to describe the editorial and jurisdictional history of the act and to insert references to entities of the domain. Being based on HTML, presentation is also offered by the same language, an aspect missing from all most notable standards for the legal domain. Furthermore, LegalHTML addresses consolidation of an act and its subsequent modifications into a single document using a tree-based representation of the original content and of its modified versions. Finally, alongside the language & ontology, we implemented a CSS stylesheet for the default rendering of LegalHTML documents and a JavaScript file imbuing documents with an API supporting TOC generation, footnote cross-references and the said point-in-time visualization of consolidated legal acts.

Domain adaptation with contrastive and adversarial oriented transferable semantic augmentation

Recent studies indicate that explicit domain adaptation methods that directly align the source and target domains will distort the original target feature distributions and cause misalignment. Implicit domain adaptation using feature augmentation bridges the domain gap and mitigates the above problems. However, such methods have the following challenges: (1) insufficient intrinsic discrimination of augmented features may cause negative transfer; (2) most methods ignore learning the domain-invariant features between source and augmented samples. To address these issues, we propose a Contrastive and Adversarial oriented transferable semantic Augmentation Domain Adaptation method (CAADA). We embed the augmentation process into our contrastive loss implicitly, minimizing it to increase the similarity between each augmented sample and the source centroid, thereby improving the augmented features’ intrinsic discriminativeness. Moreover, the prediction distribution difference between the source and augmented samples is used as the loss, and adversarial learning is employed between the classifier and feature extractor to refine domain-invariant features. The proposed losses can be integrated into other domain adaptation methods as lightweight modules to enhance each other’s performance. Experiments indicate that our proposed model yields state-of-the-art results on benchmarks, particularly resulting in 0.7% and 0.8% improvements on Office-Home and ImageCLEF-DA, respectively.

Designing Expert-Augmented Clinical Decision Support Systems to Predict Mortality Risk in ICUs

One of the most critical infrastructures during the COVID-19 pandemic are intensive care units (ICU). ICU’s crucial task is to preserve the lives of patients and mitigate the pandemic’s impact on the population. However, most ICUs plan only one day ahead. This short-term planning becomes an obstacle during disaster situations since physicians need to decide efficiently and ensure the timely treatment of high-risk patients. Integrating machine learning (ML) systems for clinical decision support could improve this process by predicting the mortality risk of critically ill patients. Several ML approaches tackling this problem have already shown promising results. However, these systems mostly neglect the integration of explicit domain knowledge, which is crucial to ensure prediction quality and adaptability. Otherwise, black-box systems might base their decision on confounding variables and improper relationships. Following design science research, we utilize a unique dataset of patients diagnosed with SARS-CoV-2 in ICU care to design a clinical decision support system by combining ML and expert knowledge in the form of a severity score. We show that by augmenting the system with expert knowledge, its overall performance improves compared to the baseline approach.

Ambient Awareness of Who Knows What: Spontaneous Inferences of Domain Expertise

ABSTRACT Being aware of others’ expertise is essential for knowledge exchange. Browsing social media has been shown to foster such awareness, often referred to as “ambient awareness”. However, the underlying psychological mechanism has not yet been established. In a series of three preregistered experiments (total N = 445), we tested the notions that individuals make inferences about others’ domains of expertise when being exposed to domain-implying social media posts, and that they do so spontaneously – that is, unintentionally and efficiently. Regardless of whether individuals were carefully reading, browsing through, or memorizing domain-implying posts, and whether they had an explicit goal to form an impression or not, we found consistent evidence that they made inferences about others’ expertise domains, as evidenced by responses in a (false) probe recognition paradigm and explicit domain identification (Experiments 1 − 3). We also showed that they did so unintentionally (Experiment 2) and efficiently (Experiment 3). Spontaneous domain inferences are a plausible psychological mechanism for how ambient awareness of who knows what is achieved through regular social media use. Our findings provide causal evidence for the proposed mechanism underlying the link between social media use and expertise localization and highlight the role of social media posts in this process.

Stable Learning via Sparse Variable Independence

The problem of covariate-shift generalization has attracted intensive research attention. Previous stable learning algorithms employ sample reweighting schemes to decorrelate the covariates when there is no explicit domain information about training data. However, with finite samples, it is difficult to achieve the desirable weights that ensure perfect independence to get rid of the unstable variables. Besides, decorrelating within stable variables may bring about high variance of learned models because of the over-reduced effective sample size. A tremendous sample size is required for these algorithms to work. In this paper, with theoretical justification, we propose SVI (Sparse Variable Independence) for the covariate-shift generalization problem. We introduce sparsity constraint to compensate for the imperfectness of sample reweighting under the finite-sample setting in previous methods. Furthermore, we organically combine independence-based sample reweighting and sparsity-based variable selection in an iterative way to avoid decorrelating within stable variables, increasing the effective sample size to alleviate variance inflation. Experiments on both synthetic and real-world datasets demonstrate the improvement of covariate-shift generalization performance brought by SVI.

Wav2vec‐MoE: An unsupervised pre‐training and adaptation method for multi‐accent ASR

AbstractIn real life, either the subjective factors of speakers or the objective environment degrades the performance of automatic speech recognition (ASR). This study focuses on one of the subjective factors, accented speech, and attempts to build a multi‐accent ASR system to solve the degradation caused by different accents, one of whose characteristic is the low resource. To deal with the challenge of the low‐resource data and the different speech styles, a wav2vec‐MoE (mixture of experts) is proposed to adapt the wav2vec 2.0 for multi‐accent ASR. In the wav2vec‐MoE, a domain MoE is developed by introducing pseudo‐domain information in the pre‐training stage, where the domain denotes a collection of speech varied by the same influence factors. The MoE is trained with two strategies according to the proposed domain mismatch assessment between unlabeled speech and target speech, without requiring any explicit domain information. Experiments show that the wav2vec‐MoE achieves a 14.69% relative word error rate reduction (WERR) on the AESRC2020 accent dataset and an 8.79% relative WERR on the Common Voice English dataset.

An Ontology-based Engineering system to support aircraft manufacturing system design

During the conceptual design phase of an aircraft manufacturing system, different industrial scenarios need to be evaluated against performance indicators in a collaborative engineering process. Domain experts’ knowledge and the motivations for decision-making is a crucial asset for enterprises which is challenging to be captured and capitalised. Ontology-based Engineering (OBE) systems emerge as a new generation of Knowledge-based Engineering techniques with advancements of ontology engineering methods and computer science technologies. Ontologies enable to capture both explicit and implicit domain knowledge from historical records and domain experts. These Ontology-based Engineering systems can stand highly complex collaborative design processes involving multidisciplinary stakeholders and various digital tools. This paper proposes a tradespace framework with Ontology-based Engineering features included on top of existing Model-Based System Engineering and interoperability capabilities. These additional Ontology-based Engineering features reuse formalised knowledge via knowledge graph technologies and generative algorithms, changing the cognitive process from the designer, to an automatic process which generates design alternatives for the designer. The tradespace framework is demonstrated in a case study to design the aircraft fuselage orbital joint process, helping the designer to take better strategic decisions at conceptual phase and proving to be an advantageous paradigm for the design process.

Barkhausen noise from formation of360∘domain walls in disordered permalloy thin films

Micromagnetic simulations reveal formation of 360${}^{\ensuremath{\circ}}$ domain walls in disordered, thin permalloy films during the field-driven magnetic reversal process. With stronger disorder, the reversal process is increasingly governed by discontinuous jumps, resulting in Barkhausen noise without explicit domain wall movement.

Exploring Implicit Domain-Invariant Features for Domain Adaptive Object Detection

Recent researches have made a great progress in domain adaptive object detectors. These detectors aim to learn explicit domain-invariant features by adversarially mitigating domain divergence and simultaneously optimizing source risks. However, an inherent problem is that they ignore the informative knowledge implied in domain-specific features, which is recognized as implicit domain-invariant feature. This is mainly caused by the multimode structure underlying target distribution, characterized by various scales and categories of objects in target images. To solve that, we propose the Implicit Domain-invariant Faster R-CNN (IDF) by using non-adversarial domain discriminator, dual attention mechanism and selective feature perception. This idea is implemented on the Faster R-CNN backbone, but with an improved architecture of two branches, i.e. domain-invariant branch and domain-specific branch. The former can clearly learn explicit domain adaptive features w.r.t. easy samples, while the latter aims to learn implicit domain-invariant features w.r.t. hard samples. Experiments on numerous benchmark datasets, including the Cityscapes, Foggy Cityscapes, KITTI and SIM10K, show the superiority of our IDF over other state-of-the-art domain adaptive object detectors. The demo code is released in <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/sea123321/IDF</uri> .

Frequency domain transfer function based multimodal design method for seismic mitigation of building structures

Multiple tuned mass dampers (MTMD) are widely used to mitigate structural seismic response. With the emergence of more and more high-rise buildings, various MTMD design methods applicable to multi degree of freedom (MDOF) model have been developed. However, the existing MTMD design methods for MDOF models often involve complex optimization algorithms. In order to simplify the MTMD design while effectively reducing the seismic response of the structure, this paper proposes an explicit frequency domain transfer function, theoretically showing the mechanism of decomposing the seismic mitigation of MDOF model into simultaneous seismic mitigation of multiple modes, and developed a theoretical design method of MTMD, so that the dynamic parameters and position of each TMD can be directly obtained based on the developed explicit design formula. A benchmark model was used as an example to verify the seismic mitigation performance. It is shown that the developed multimodal control MTMD design method achieves 22.25% reduction rate of acceleration and 18.5% reduction rate of displacement with 3% mass ratio.

Stable attenuation-compensated reverse time migration and its application to land seismic data

Intrinsic attenuation of the earth causes energy loss and phase distortion in seismic wave propagation. To obtain high-resolution imaging results, these negative effects must be considered during reverse time migration (RTM). We can easily implement attenuation-compensated RTM using the constant Q viscoacoustic wave equation with decoupled amplitude attenuation and phase dispersion terms. However, the nonphysical amplitude-compensation process will inevitably amplify the high-frequency noise in the wavefield in an exponential form, causing the numerical simulation to become unstable. This is due to the fact that the amplitude of the compensation grows exponentially with frequency. In order to achieve stable attenuation-compensated RTM, we modify the analytic expression of the attenuation compensation extrapolation operator and make it only compensate for amplitude loss within the effective frequency band. Based on this modified analytic formula, we then derive an explicit time-space domain attenuation compensation extrapolation operator. Finally, the implementation procedure of stable attenuation-compensated RTM is presented. In addition to being simple to implement, the newly proposed attenuation-compensated extrapolation operator is superior to the conventional low-pass filter in suppressing random noise, which will further improve the imaging resolution. We use two synthetic and one land seismic datasets to verify the stability and effectiveness of the proposed attenuation-compensated RTM in improving imaging resolution in viscous media.

A simplified method to consider hydrodynamic effect in oblique vessel-bridge collisions

The hydrodynamic effect in vessel-bridge collisions is either directly omitted or considered by the constant added mass (CAM) method in current studies. However, the high-resolution fluid-structure interaction (FSI) simulations indicate that the hydrodynamic effect plays an important role and cannot be adequately captured by the CAM method, particularly for oblique vessel-bridge collisions. On the other hand, the FSI simulations are time-consuming and have a high requirement for computational resources. For these reasons, a simplified analysis method was developed in this study to consider the hydrodynamic effect of vessel collision with bridge structures. In the developed method, mechanics of full 6-DoF ship maneuvering were established using several matrices to replace the explicit fluid domain of the FSI models, while the added mass method was adopted for the impacted pier to account for the hydrodynamic effect. Several collision scenarios were performed to compare the developed method with the FSI simulation and the CAM analysis. It was found that the proposed method is superior to the CAM method in terms of considering the hydrodynamic effect of vessel-bridge collisions, and its required computational cost can be significantly saved in comparison with the FSI model. From the perspective of calculation efficiency and accuracy, the proposed method is more suitable for the analysis and design of vessel-bridge collision in engineering practice than the FSI simulation and the CAM analysis.