154092 research outputs found
Sort by
Perturbation-Aware Distributionally Robust Optimization for Inverse Problems
No full textThis paper builds on classical distributionally robust optimization techniques to construct a comprehensive framework that can be used for solving inverse problems. Given an estimated distribution of inputs in X and outputs in Y, an ambiguity set is constructed by collecting all the perturbations that belong to a prescribed set K and are inside an entropy-regularized Wasserstein ball. By finding the worst-case reconstruction within K one can produce reconstructions that are robust with respect to various types of perturbations: X-robustness, Y|X-robustness and, more general, targeted robustness depending on noise type, imperfect forward operators and noise anisotropies. After defining the general robust optimization problem, we derive its (weak) dual formulation and we use it to design an efficient algorithm. Finally, we demonstrate the effectiveness of our general framework to solve matrix inversion and deconvolution problems defining K as the set of multivariate Gaussian perturbations in Y|X
Simulating workload reduction with an AI-based prostate cancer detection pathway using a prediction uncertainty metric
Get PDFObjectives: This study compared two uncertainty quantification (UQ) metrics to rule out prostate MRI scans with a high-confidence artificial intelligence (AI) prediction and investigated the resulting potential radiologist’s workload reduction in a clinically significant prostate cancer (csPCa) detection pathway. Materials and methods: This retrospective study utilized 1612 MRI scans from three institutes for csPCa (Gleason Grade Group ≥ 2) assessment. We compared the standard diagnostic pathway (radiologist reading) to an AI-based rule-out pathway in terms of efficacy and accuracy in diagnosing csPCa. In the rule-out pathway, 15 AI submodels (trained on 7756 cases) diagnosed each MRI scan, and any prediction deemed uncertain was referred to a radiologist for reading. We compared the mean (meanUQ) and variability (varUQ) of predictions using the DeLong test on the area under the receiver operating characteristic curves (AUROC). The level of workload reduction of the best UQ method was determined based on a maintained sensitivity at non-inferior specificity using the margins 0.05 and 0.10. Results: The workload reduction of the proposed pathway was institute-specific: up to 20% at a 0.10 non-inferiority margin (p < 0.05) and non-significant workload reduction at a 0.05 margin. VarUQ-based rule out gave higher but non-significant AUROC scores than meanUQ in certain selected cases (+0.05 AUROC, p > 0.05). Conclusion: MeanUQ and varUQ showed promise in AI-based rule-out csPCa detection. Using varUQ in an AI-based csPCa detection pathway could reduce the number of scans radiologists need to read. The varying performance of the UQ rule-out indicates the need for institute-specific UQ thresholds. Key Points: Question AI can autonomously assess prostate MRI scans with high certainty at a non-inferior performance compared to radiologists, potentially reducing the workload of radiologists. Findings The optimal ratio of AI-model and radiologist readings is institute-dependent and requires calibration. Clinical relevance Semi-autonomous AI-based prostate cancer detection with variational UQ scores shows promise in reducing the number of scans radiologists need to read.</p
Integrated models of nutrient dynamics in lake and reservoir watersheds:A systematic review and integrated modelling decision pathway
Get PDFEutrophication of inland water bodies is a serious environmental threat. This review explores current integrated models for lake and reservoir ecosystems that focus on nutrient dynamics at a catchment scale. Many studies applied either watershed or lake/reservoir models, however, 49 studies were finally selected that combined both. We derived a list of 21 watershed models, 23 lake/reservoir models, and 6 hybrid models in different sets of combinations, with a range of objectives (e.g. understanding the natural processes, predicting, and analysing climate change and land-use scenarios, or evaluating the different management options). Some integrated models had multiple applications whereas others were only applied once, with an uneven global geographical distribution. To aid model selection by future users, we present a support tool discriminating the models by their features and application fields. This study encourages the development of open-source tools aiding interdisciplinary collaborations and further research in the field of integrated modelling.</p
Just-in-Time in situ pre-operative three-dimensional visualisation for anatomical lung surgery
No full textIntroduction: context and hypothesis/aimsFor patients with local non-small cell lung cancer <2cm with poor pulmonary function, minimally-invasive surgical removal of a single or multiple lung segments (anatomical lung resection) might be considered as a treatment option.The anatomical lung resection is, however, a complex and low-volume procedure. The use of three-dimensional (3D) models is proven to be of added value for these procedures and can potentially decrease intraoperative blood loss, operative time and patient complications.Commercial services may provide preoperative models, yet these services are costly and require transfer of sensitive patient data to external entities.This study aimed to improve patient informed consent and facilitate pre-operative planning of anatomical lung resections with 3D models compared to standard 2D contrast-enhanced CT scans.We hypothesize that 3D models lead to changes in surgical plans and improve surgical anatomy knowledge. Methods and results: description of the methods used/study design/data collection. Presentation of the results addressing the study hypothesis/aimsAn in-house CT-scan segmentation protocol using Mimics (Materialise N.V., Leuven, Belgium) software was developed and implemented at Medisch Spectrum Twente (MST), the Netherlands. Patients with stage 1A1-2 lung cancer who underwent a segmentectomy from January 2023 to August 2024 were included in the study. Informed consent forprospective patients was obtained after local ethics approval. Two surgeons filled in questionnaires to estimate the change of surgical plan and model quality (Kirkpatrick level 2 learning and level 3 behaviour change).In total, 14 patients were included to make 3D models. Questionnaires revealed that surgeons felt comfortable creating a preoperative plan using the 3D models. The quality of the 3D models was perceived to be good or excellent in all but one case. The use of 3D models changed 32% of all surgical plans compared to 2D CT. Out of these 14 patients, 3patients received preoperative 3D models prospectively. For these patients, all relevant structures identified preoperatively in the 3D model were identified intra-operatively. In all cases, the just-in-time 3D models showed to have added value in increasing understanding of patient-specific anatomy enabling the surgeons to more thoroughly preparefor each case.Discussion of the impact/outcome, and novelty of the ResearchUsing in-situ preoperative 3D models created by technical medical doctors leads to a significant change in surgical plan and an increased understanding of the patient’s specific anatomical variations, potentially leading to more accurate and safer anatomical resections in the future. Because of low-volume surgery, further research must investigate the clinical benefits of 3D-guided lung segmentectomies (Kirkpatrick Level 4).Keywords3D visualisation, in situ simulation, minimally-invasive surgery, Cardio-thoracic surgeryReferences/Acknowledgements1. Kato H, Oizumi H, Suzuki J, Suzuki K, Takamori S, Kato H, et al. Indications and technical details of sublobarresections for small-sized lung cancers based on tumor characteristics. Mini-invasive Surg 2021;5:5. 2021-02-03;5(0).2. Yotsukura M, Okubo Y, Yoshida Y, Nakagawa K, Watanabe S-i. Indocyanine green imaging for pulmonarysegmentectomy. JTCVS Techniques. 2021/04/01;6
Towards an Ontology of Type-Level Phenomena for System Modeling
No full textIn this paper, we propose a well-founded domain-independent system ontology based on the Unified Foundational Ontology (UFO) and the associated Multi-Level Theory (MLT). In our ontology, (composite) system types are designed by defining component types, connection types, and other type-level phenomena. We discuss how the proposed ontology can be used to guide the modeling of systems in specific domains. We position our work with respect to a number of ontologies in the literature that address system-related notions, as well as with respect to system modeling approaches.</p
Towards a Celeste AI Framework:Agent-free Automated 2D Level Generation for Multidirectional Platformers
No full textWe present a procedural content generation (PCG) pipeline for Celeste, a complex 2D platformer with horizontal and vertical movement and with limited prior AI framework development. Our approach utilizes a Markov Chain-based model to capture the game's unique structural and gameplay elements, generating playable levels that adhere to Celeste's design principles. We implemented post-processing steps to enhance playability and strategically place game elements. Our evaluation metrics focused on playability and interestingness, with results that indicate success in replicating the desired gameplay experience for beginner players. The evaluation involved 12 players of different skill levels, providing insight into the effectiveness of our generated content. Although some limitations were observed, such as occasional lack of creativity and difficulty in controlling challenge levels, our pipeline demonstrates promise as a foundation for a Celeste AI framework. This study contributes to the broader field of PCG for complex platformers and opens avenues for level generation in which agent-based evaluation is not feasible.</p
Median of Forests for Robust Density Estimation
No full textRobust density estimation refers to the consistent estimation of the density function even when the data is contaminated by outliers. We find that existing forest density estimation at a certain point is inherently resistant to the outliers outside the cells containing the point, which we call \textit{non-local outliers}, but not resistant to the rest \textit{local outliers}. To achieve robustness against all outliers, we propose an ensemble learning algorithm called \textit{medians of forests for robust density estimation} (\textit{MFRDE}), which adopts a pointwise median operation on forest density estimators fitted on subsampled datasets. Compared to existing robust kernel-based methods, MFRDE enables us to choose larger subsampling sizes, sacrificing less accuracy for density estimation while achieving robustness. On the theoretical side, we introduce the local outlier exponent to quantify the number of local outliers. Under this exponent, we show that even if the number of outliers reaches a certain polynomial order in the sample size, MFRDE is able to achieve almost the same convergence rate as the same algorithm on uncontaminated data, whereas robust kernel-based methods fail. On the practical side, real data experiments show that MFRDE outperforms existing robust kernel-based methods. Moreover, we apply MFRDE to anomaly detection to showcase a further application
On the Robustness of Kernel Ridge Regression Using the Cauchy Loss Function
No full textRobust regression aims to develop methods for estimating an unknown regression function in the presence of outliers, heavy-tailed distributions, or contaminated data, which can severely impact performance. Most existing theoretical results in robust regression assume that the noise has a finite absolute mean, an assumption violated by certain distributions, such as Cauchy and some Pareto noise. In this paper, we introduce a generalized Cauchy noise framework that accommodates all noise distributions with finite moments of any order, even when the absolute mean is infinite. Within this framework, we study the \textit{kernel Cauchy ridge regressor} (\textit{KCRR}), which minimizes a regularized empirical Cauchy risk to achieve robustness. To derive the -risk bound for KCRR, we establish a connection between the excess Cauchy risk and -risk for sufficiently large scale parameters of the Cauchy loss, which reveals that these two risks are equivalent. Furthermore, under the assumption that the regression function satisfies H\"older smoothness, we derive excess Cauchy risk bounds for KCRR, showing improved performance as the scale parameter decreases. By considering the twofold effect of the scale parameter on the excess Cauchy risk and its equivalence with the -risk, we establish the almost minimax-optimal convergence rate for KCRR in terms of -risk, highlighting the robustness of the Cauchy loss in handling various types of noise. Finally, we validate the effectiveness of KCRR through experiments on both synthetic and real-world datasets under diverse noise corruption scenarios
On the expressivity of deep Heaviside networks
No full textWe show that deep Heaviside networks (DHNs) have limited expressiveness but that this can be overcome by including either skip connections or neurons with linear activation. We provide lower and upper bounds for the Vapnik-Chervonenkis (VC) dimensions and approximation rates of these network classes. As an application, we derive statistical convergence rates for DHN fits in the nonparametric regression model
Ontological Analysis of Advanced Capability Modeling in ArchiMate:A First Step Towards Language Revision
Get PDFIn order to support capability management, the field of Enterprise Architecture proposes methods and notations to model enterprises and their capabilities. ArchiMate is one of such notations and includes constructs to support capability mapping and other capability management tasks. However, the notation lacks some fine-grained distinctions that are required to understand intricate phenomena involving capabilities, including capability interaction and the emergence of capabilities. In this work, we perform an ontological analysis of the language’s support for capability modeling based on a well-founded ontology of capabilities aligned with the Unified Foundational Ontology (UFO). Through this ontological analysis, we identify some issues outlining possible improvements for ArchiMate. This is a first step towards language redesign, which may include the proposal of language patterns and/or the revision of language constructs.</p