Decision Rules Research Articles

Explaining deep learning for ECG analysis: Building blocks for auditing and knowledge discovery

Deep neural networks have become increasingly popular for analyzing ECG data because of their ability to accurately identify cardiac conditions and hidden clinical factors. However, the lack of transparency due to the black box nature of these models is a common concern. To address this issue, explainable AI (XAI) methods can be employed. In this study, we present a comprehensive analysis of post-hoc XAI methods, investigating the glocal (aggregated local attributions over multiple samples) and global (concept based XAI) perspectives. We have established a set of sanity checks to identify saliency as the most sensible attribution method. We provide a dataset-wide analysis across entire patient subgroups, which goes beyond anecdotal evidence, to establish the first quantitative evidence for the alignment of model behavior with cardiologists’ decision rules. Furthermore, we demonstrate how these XAI techniques can be utilized for knowledge discovery, such as identifying subtypes of myocardial infarction. We believe that these proposed methods can serve as building blocks for a complementary assessment of the internal validity during a certification process, as well as for knowledge discovery in the field of ECG analysis.

An Ensemble Approach To Face Recognition In Access Control Systems

An Ensemble Approach To Face Recognition In Access Control Systems

Interval-valued intuitionistic fuzzy three-way conflict analysis based on cumulative prospect theory

This paper makes a significant contribution to the field of conflict analysis by introducing a novel Interval-Valued Intuitionistic Fuzzy Three-Way Conflict Analysis (IVIFTWCA) method, which is anchored in cumulative prospect theory. The method’s key innovation lies in its use of interval-valued intuitionistic fuzzy numbers to represent an agent’s stance, addressing the psychological dimensions and risk tendencies of decision-makers that have been largely overlooked in previous studies. The IVIFTWCA method categorizes conflict situations into affirmative, impartial, and adverse coalitions, leveraging the evaluation of the closeness function and predefined thresholds. It incorporates a reference point, value functions and cumulative weight functions to assess risk preferences, leading to the formulation of precise decision rules and thresholds. The method’s efficacy and applicability are demonstrated through detailed examples and comparative analysis, and its exceptional performance is confirmed through a series of experiments, offering a robust framework for real-world decision-making in conflict situations.

Abstract PO2-19-11: NOBLE – Neoadjuvant Olaparib and Durvalumab for patients with BRCA-associated TripLE Negative Breast Cancer

Abstract Background The standard of care for patients with early triple-negative breast cancer (TNBC) is neoadjuvant chemotherapy followed by surgery. Recently, immune checkpoints inhibitors (ICI) added to neoadjuvant chemotherapy have shown efficacy. Among TNBC, BRCA mutations and other homologous recombination deficiency (HRD) signals are relatively frequent. As opposed to other types of breast cancer, TNBC show a high neoantigen load, a large number of tumour infiltrating lymphocytes (TILs), and a high PD-L1 expression. In BRCA/HRD breast cancer, PARP inhibitors (PARPi) exhibit a distinct anticancer activity. In addition, they show immunomodulatory effects by promoting PD-L1 expression and DNA damage, increasing neoantigen load. Chemotherapy-free regimens combining PARPi and ICI result in promising synergistic activity as shown in the metastatic setting, supporting investigation in the neoadjuvant setting. Trial design NOBLE is an EORTC phase 2, randomized, international, multicentre, open label, clinical trial. Eligible patients will undergo central screening for tumoral BRCA/HRD (Institute Curie, 455 patients), 152 patients will be randomized 1:1 between olaparib vs olaparib and durvalumab for 4 cycles (16 weeks) of neoadjuvant treatment. Thereafter, patients will undergo surgery. All patients will be followed-up to 2 years (y) after surgery. The trial will be conducted in 3 steps (screening/randomization): Step 1 – Feasibility (60/20): to assess the central testing failure rate, the turn-around time for central testing and the prevalence of tBRCA/HRD; Step 2 (193/64) – Futility; and Step 3 (202/152) – Expansion of recruitment. The primary endpoint is the rate of pathologic complete response (pCR) at the time of surgery, defined as ypT0/is ypN0. Secondary endpoints are response based on RECIST v1.1, surgery rate, 2-y event-free-survival, 2-y overall survival, safety, and quality of life. The study is expected to be activated in early autumn 2023 and the primary endpoint will be available approximately 5 years after the screening of the first patient. Main inclusion criteria Histologically confirmed primary TNBC, defined as: - ER and PgR ≤ 10% - HER2-negative per ASCO CAP guidelines Stage T1c-T2 N0-N1 M0 BRCA mutation and/or HRD based on methylation as determined by central testing ECOG status 0-1 Written informed consent Age ≥18 years No prior systemic therapy nor surgery for the current Main exclusion criteria Previous treatment with a PARPi or an immune-checkpoint inhibitor History of previous invasive breast cancer Bilateral and/or multifocal and/or multicentric BC Autoimmune or inflammatory disease Specific aims The aim of this study is to provide a safe and effective neoadjuvant treatment free of chemotherapy to a distinct subpopulation among patients with TNBC. We will evaluate the pCR after neoadjuvant treatment with olaparib alone and with olaparib in combination with durvalumab in patients with TNBC showing a BRCA1/2 mutation or HRD based on methylation. Statistical methods The sample size was calculated based on the minimax Simon two-stage design including an interim analysis for futility of the primary endpoint. Decision rules are pre-specified in the study protocol. Each treatment arm will be analyzed independently. There will be no formal comparison between the two experimental treatment arms. A minimization technique will be used for random treatment allocation stratifying for the type of HRD alteration, PD-L1 expression level and for TILs. Accrual We expect to start recruiting in September 2023. We aim to screen 455 patients to randomize 152 patients. Contact information for people with a specific interest in the trial Etienne Brain (study coordinator): etienne.brain@curie.fr Anne-Sophie Hamy-Petit (study co-coordinator): anne-sophie.hamy-petit@curie.fr Citation Format: Emanuel Buhrer, Eleni Xenophontos, Anne-Sophie Hamy-Petit, Jose Casas-Martin, Céline Callens, Thomas Meyskens, Coralie Poncet, Anastasia Lanzi, Etienne Brain. NOBLE – Neoadjuvant Olaparib and Durvalumab for patients with BRCA-associated TripLE Negative Breast Cancer [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO2-19-11.

Computed tomography pulmonary angiogram ordering, adherence to decision rules and yield in the emergency department: An observational study.

Pulmonary embolism (PE) frequently requires diagnosis through CT pulmonary angiogram (CTPA). Appropriate application of evidence-based clinical decision tools can reduce unnecessary CTPAs. This study assessed adherence to and the efficacy of various aspects of the Queensland Health suspected PE diagnostic pathway, including Wells score, PE rule out criteria (PERC) and age-adjusted D-dimer interpretation. Retrospective study of CTPAs ordered from 1 January to 30 April 2023 in a tertiary and urban ED in Southeast Queensland. Data on clinical variables, D-dimer and CTPA results were collected through medical record and radiology database review. Descriptive analyses were used to determine adherence to Queensland guidelines and performance of D-dimer interpretation tools (including comparison of age-adjusted PE with a new pre-test probability [PTP]-based model using D-dimer cut-off <1000 ng/mL for Wells score ≤4 and 500 ng/mL for Wells score 4.5-6). A total of 573 CTPAs were available for analysis with a 12.4% (95% confidence interval 10.0-15.4) diagnostic yield. Stratification by Wells score showed yields of 4.0%, 18.5% and 41% for low-, moderate- and high-risk patients, respectively. Twenty-five patients with low-PTP who received CTPA could have been excluded with the PERC rule. Age-adjusted D-dimer interpretation may have prevented 26 CTPAs with no false negatives, whereas PTP approach may have prevented 128 CTPAs with four false negatives. Guideline adherence can be improved, and adherence to existing clinical decision tools may reduce unnecessary CTPA ordering and increase diagnostic yield. The use of the age-adjusted D-dimer had good sensitivity, whereas the new PTP approach will require further prospective research.

SUSTAINABILITY TRANSPARENCY IN SOVEREIGN WEALTH FUNDS: BENCHMARK AND CASE STUDY APPROACH

The COVID-19 pandemic and the current food, energy, and safety crisis caused by the war in Ukraine strongly affected the existing gap in Sustainable Development Goals funding. Restoring global sustainability progress and the 2030 Agenda requires reshaping responsible investment architecture and rethinking the role of sovereign wealth funds as the largest investors in this process.The current study explores the ways in which the most transparent sovereign wealth funds act towards sustainability values and incorporate environmental, social, governance, and ethical criteria (ESGE) into their investment strategy, portfolio, and reporting. Also, 16 high-ranked funds by the Sustainability Transparency Index (Cavagnetto et al., 2022) were investigated by using the case study and benchmark analysis. Study results demonstrate the similarity of all funds disclosure within some questionnaire criteria like sustainability policy, scoring (voting) rules for decision-making, ESGE criteria, stakeholder engagements, and assurance of the financial information disclosed. Nevertheless, qualitative differences were identified between the sustainability transparency criteria of A and B-ranked funds. There is also huge variation in the implementation of funds’ ethical criteria, sustainability coherence, relevance of sustainability goals, disclosure of specific climate-related goals, and foundation membership criteria.Policy implications derived from the differences in sovereign funds' benchmarks are aimed at enhancing the sustainability context in funds’ operating frameworks, developing impact measurement metrics, prioritizing the most material stakeholder requests, sustainability verification procedures, and standardizing funds’ sustainability reporting. Specific channels, instruments, and regulatory measures towards greater funds’ sustainability transparency and their involvement in bridging the sustainable development funding gap may be the subject of further investigation.

Fuzzy three-way rule learning and its classification methods

Rules play a crucial role in classification tasks, driving the advancement of artificial intelligence. However, how to improve the interpretability of extracted rules while ensuring the performance of classification tasks is always a challenge, owing to the diversity of data types. Since three-way decision rules derive and explain from positive and negative aspects and provide more detailed information than general rules, this article explores fuzzy three-way rule learning from the perspective of two-way granular reduct by taking the FCA-based granular computing method as a framework. Specifically, we first present the object-induced fuzzy three-way granular rules and the object-induced two-way fuzzy three-way rules. Then, the fuzzy three-way rule-based dynamic updating method (FTRDUM) and the weight-based voting method are proposed to improve the classification performance. Finally, to illustrate the effectiveness of FTRDUM, some numerical experiments are conducted. The results show the superiority of the proposed algorithm in classification accuracy.

Reinforcement Learning from AI Feedback A Review

Reinforcement Learning from AI Feedback (RLAIF) is a big step forward compared to Reinforcement Learning from Human Feedback (RLHF). It's especially useful for large language models like GPT-4. RLAIF is better because it can handle more data at scale and is more efficient. It uses AI-generated feedback instead of human feedback. This shift to AI-generated feedback enhances the efficiency and speed of training AI systems. Additionally, RLAIF optimizes the AI's ability to align with desired outcomes, although it may not directly improve understanding human preferences. RLAIF uses a Preference Model (PM) that follows constitutional principles. This ensures that AI responses are ethical, safe, and high -quality. The constitution sets rules for AI decision-making. It makes sure AI follows ethical and social standards. This is important as AI keeps evolving. RLAIF is moving towards an automated, moral feedback system focusing on responsible AI governance and ethical guidelines.

Fraud risk assessment in car insurance using claims graph features in machine learning

The article proposes a process for claims assessment in car insurance, which makes it possible to calculate the fraud rate on the annual set of claims using a reduced set of attributes and graph vertex properties. This approach improves the security of insurance companies’ assets against fraudulent attacks. A method for constructing a claims graph and extracting additional features from it for evaluation is described. It is shown that in order to build a graph, it is not necessary to have data on the connection of the claim participants. Two tests were carried out on a real opensource datasets with labelling of fraudulent cases. The results of the first one show the increase in classification metrics when using attributes obtained from the graph. The application of the proposed approach resulted in doubling the area under the Precision-Recall curve. The experimental results demonstrated high quality metrics for fraud detection, with a Recall rate of 83.33% and a Specificity rate of 91.05%. The second test confirmed the possibility of determining the insurance fraud level based on decision rule, which includes the condition of claims being connected to each other. The rule is able to detect claim groups with a high concentration of fraud, in which every second participant is a fraudster.

Incremental Green Investment Rule Induction Using Intelligent Rough Sets from an Energy Perspective

The United Nations Sustainable Development Goals (SDGs) agenda has stated the importance of green investment. Energy-related green investment involves intricate economic behavior and ecological objectives. Green investment definitely requires agile decisions, e.g., rule-based decisions, to respond to changes outside the country. The identification of significant rules with numerous result features and the assurance of the stability and robustness of the rules in decision-making are crucial for green energy investment. The rough set (RS) methodology works well for processing qualitative data that are difficult to examine with traditional statistical methods in order to induce decision rules. The RS methodology starts with the analysis of the limits of discernibility of a subset of objects belonging to the domain to induce rules. However, traditional RS methods cannot incrementally generate rules with outcome features when new objects are added, which frequently occurs in green energy investment with the inclusion of big data. In this paper, an intelligent RS approach is proposed. This approach effectively identifies the rules that either stay the same or are altered based on four classified cases after a new object is introduced; it is novel because it can deal with a complicated investment environment by imposing multiple outcome features, specifically when it is required to flexibly extract new decision rules via adding new data sets.

Integrating decision modeling and machine learning to inform treatment stratification.

There is increasing interest in moving away from "one size fits all (OSFA)" approaches toward stratifying treatment decisions. Understanding how expected effectiveness and cost-effectiveness varies with patient covariates is a key aspect of stratified decision making. Recently proposed machine learning (ML) methods can learn heterogeneity in outcomes without pre-specifying subgroups or functional forms, enabling the construction of decision rules ('policies') that map individual covariates into a treatment decision. However, these methods do not yet integrate ML estimates into a decision modeling framework in order to reflect long-term policy-relevant outcomes and synthesize information from multiple sources. In this paper, we propose a method to integrate ML and decision modeling, when individual patient data is available to estimate treatment-specific survival time. We also propose a novel implementation of policy tree algorithms to define subgroups using decision model output. We demonstrate these methods using the SPRINT (Systolic Blood Pressure Intervention Trial), comparing outcomes for "standard" and "intensive" blood pressure targets. We find that including ML into a decision model can impact the estimate of incremental net health benefit (INHB) for OSFA policies. We also find evidence that stratifying treatment using subgroups defined by a tree-based algorithm can increase the estimates of the INHB.

HiSOMA: A hierarchical multi-agent model integrating self-organizing neural networks with multi-agent deep reinforcement learning

Multi-agent deep reinforcement learning (MADRL) has shown remarkable advancements in the past decade. However, most current MADRL models focus on task-specific short-horizon problems involving a small number of agents, limiting their applicability to long-horizon planning in complex environments. Hierarchical multi-agent models offer a promising solution by organizing agents into different levels, effectively addressing tasks with varying planning horizons. However, these models often face constraints related to the number of agents or levels of hierarchies. This paper introduces HiSOMA, a novel hierarchical multi-agent model designed to handle long-horizon, multi-agent, multi-task decision-making problems. The top-level controller, FALCON, is modeled as a class of self-organizing neural networks (SONN), designed to learn high-level decision rules as internal cognitive codes to modulate middle-level controllers in a fast and incremental manner. The middle-level controllers, MADRL models, in turn receive modulatory signals from the higher level and regulate bottom-level controllers, which learn individual action policies generating primitive actions and interacting directly with the environment. Extensive experiments across different levels of the hierarchical model demonstrate HiSOMA’s efficiency in tackling challenging long-horizon problems, surpassing a number of non-hierarchical MADRL approaches. Moreover, its modular design allows for extension into deeper hierarchies and application to more complex tasks with heterogeneous controllers. Demonstration videos and codes can be found on our project web page: https://smu-ncc.github.io.

РОЗРОБКА ЕКСПЕРТНОЇ СИСТЕМИ ДЛЯ ВИБОРУ СТИЛЮ ОДЯГУ НА ОСНОВІ KANSEI ENGINEERING

This research is devoted to implementation of the expert systems for rapid change in production of women's outerwear. Factorial analysis and the cluster analysis are used for the structuring of the subject environment. Analysis of the emotional component of the garment on the basis of KE can be achieved through cluster analysis. Such an approach will identify models of clothes based on perception and emotional needs of the consumer. Thus, the main goal of the study was achieved by forming decision-making rules aimed at solving the sub-task of selecting ready-to-wear models based on the Kansei Engineering methodology. The analysis and identification of the formed database of photos of ready-made clothing models was carried out by means of cluster analysis of intergroup relations using the Euclidean distance, which allows for the identification of conditional groups of relatedness of products by design attributes. The obtained final centers of the clusters together with the information about the color solution of the clothing models allowed us to build a product model of an expert system for choosing clothing that meets the formulated wishes of the target consumer. The object of the study is the women's fashion dresses in spring-summer 2023 season. Thus, a general collection was formed which amounted to 66 photos of fashion dresses for subsequent questionnaire. Work with the software product takes place in the form of a dialogue in the form of consecutive system questions and user answers. The developed system can be used both for the selection of ready-made clothes (for example, in stores, including online stores), and for the selection of a prototype for the development of a new model of clothing that meets the aesthetic needs of the consumers.

Three-way decision method based on triangular norms in incomplete information systems and its applications in medical diagnosis

Three-way decision, as an extension of traditional two-way decision, was proposed by Yao in 2009, which can effectively avoid unnecessary losses caused by incorrect decisions in the decision-making process. Meanwhile, incomplete hybrid information systems represent the database of the relationship between objects and attributes, which refers to a system with multiple data and missing data. In this paper, based on the fact that many existing literatures involving incomplete hybrid information systems did not fully consider the impact of different conditional attributes on decision attributes and lacked effective aggregation methods to integrate weights and distances, we propose a new three-way decision method to deal with incomplete hybrid information systems with the help of triangular norms. First, in incomplete hybrid information systems, we redefine the distance between two objects based on conditional attributes and give the calculation formula of different data attributes in conditional attributes. At the same time, we define a new weight calculation method based on the relationship between conditional attributes and decision attributes. And then, by the distance between different conditional attributes and their corresponding weights, the hybrid distance is obtained using triangular norms. Furthermore, we use the hybrid distance to get the tolerance relation on the target set of an incomplete hybrid information system, thereby, using this tolerance relation, we get a new decision theoretic rough set model and the corresponding decision rules. Finally, by two experiments involving medical diagnosis, we demonstrate that our model has better classification ability, lower misclassification rate, and better stability compared to other corresponding models, thereby confirming that the proposed model provides a new and effective method for handling incomplete hybrid information systems in practical applications.

Real-time Gait Event Tracker (RGET): An Innovative Approach to Digital Human Modeling through Integration of Inertial Measurement Unit and Dynamic Feature Extraction

The advancement of digital human modeling technology underscores the importance of accurately representing human characteristics and behaviors using motion capture technologies. This paper introduces the Real-time Gait Event Tracker (RGET), an advanced method aimed at precisely detecting gait events via an Inertial Measurement Unit (IMU), thereby supporting high-precision digital human modeling. RGET integrates planning techniques into continuous states and action spaces, utilizing first-order difference functions and sliding window techniques to effectively identify four key gait events: Heel Strike (HS), Toe Off (TO), Walking Start (WS), and Walking Pause (WP). The system efficiently manages IMU gait signal data through real-time queuing techniques and extracts dynamic features within time frames using positive/negative windowed first-order difference integrations. By adopting weighted sleep time methods and adaptive threshold decision rules, RGET accurately extracts gait event features from sequences. Not only does RGET support precise gait segmentation with an F1 score of 95.9%, but it also provides an innovative method for digital human modeling and its applications.

Socioeconomic agents as active matter in nonequilibrium Sakoda-Schelling models.

How robust are socioeconomic agent-based models with respect to the details of the agents' decision rule? We tackle this question by considering an occupation model in the spirit of the Sakoda-Schelling model, historically introduced to shed light on segregation dynamics among human groups. For a large class of utility functions and decision rules, we pinpoint the nonequilibrium nature of the agent dynamics, while recovering an equilibrium-like phase separation phenomenology. Within the mean-field approximation we show how the model can be mapped, to some extent, onto an active matter field description. Finally, we consider nonreciprocal interactions between two populations and show how they can lead to nonsteady macroscopic behavior. We believe our approach provides a unifying framework to further study geography-dependent agent-based models, notably paving the way for joint consideration of population and price dynamics within a field theoretic approach.

Estimating dynamic treatment regimes for ordinal outcomes with household interference: Application in household smoking cessation.

The focus of precision medicine is on decision support, often in the form of dynamic treatment regimes, which are sequences of decision rules. At each decision point, the decision rules determine the next treatment according to the patient's baseline characteristics, the information on treatments and responses accrued by that point, and the patient's current health status, including symptom severity and other measures. However, dynamic treatment regime estimation with ordinal outcomes is rarely studied, and rarer still in the context of interference - where one patient's treatment may affect another's outcome. In this paper, we introduce the weighted proportional odds model: a regression based, approximate doubly-robust approach to single-stage dynamic treatment regime estimation for ordinal outcomes. This method also accounts for the possibility of interference between individuals sharing a household through the use of covariate balancing weights derived from joint propensity scores. Examining different types of balancing weights, we verify the approximate double robustness of weighted proportional odds model with our adjusted weights via simulation studies. We further extend weighted proportional odds model to multi-stage dynamic treatment regime estimation with household interference, namely dynamic weighted proportional odds model. Lastly, we demonstrate our proposed methodology in the analysis of longitudinal survey data from the Population Assessment of Tobacco and Health study, which motivates this work. Furthermore, considering interference, we provide optimal treatment strategies for households to achieve smoking cessation of the pair in the household.

Risk sensitive linear approximations

We propose a linear approximation to the solution of DSGE models that is sensitive to the effects of risk. If variables remain close to the approximation point in expectation, a second-order Taylor expansion to the equilibrium conditions reduces to a fixed-point problem characterized by a system of linear equations that depends on the second-order moments of the variables. The latter can be solved recursively using standard linear rational expectation methods. The resulting approximation captures the effects of risk in models with constant volatility, stochastic volatility, and GARCH effects through the intercept and/or the slopes of the decision rules. Relative to alternative approximations, our method yields approximation errors that are up to two orders of magnitude smaller, all while preserving a linear structure in the state variables. Finally, we show how to accommodate asymmetric effects from non-normal shocks within our linear approximation using information from a third-order Taylor expansion.

The Exploring the Relationship between Financial Performance Indicators, ESG, and Stock Price Returns: A Rough Set-based Bipolar Approach

Using a Rough-Set-based approach, this study investigates the connection between financial performance, Environmental, Social, and Governance (ESG) factors, and stock returns. It analyzes data from various companies and industries to examine how ESG metrics interact with financial indicators to influence stock prices. Those findings underscore the importance of considering financial and ESG factors when evaluating a company's performance and market valuation. The Rough-Set-based bipolar model comprises eight decision rules, and the DEMATEL (Decision-Making Trial and Evaluation Laboratory) analysis reveals the influence relationship among the core attributes. Furthermore, the ranking result, based on the bipolar model, is consistent with the market returns, which suggests the validity of this study.

A novel attribute reduction method with constraints on empirical risk and decision rule length

Attribute reduction is a crucial issue of rough set theory and has been applied to various fields. It aims to remove useless or redundant features from data and extract precise rules. Traditional attribute reduction methods have obvious limitations, that is, it can not guarantee the generalization ability of reduced decision rules to classification but can only guarantee the classification performance to visible objects. To overcome this issue and enhance the model classification generalization ability without increasing rules' complexity based attribute reduction, we propose a novel attribute method by employing the Structural Risk Minimization principle, which is a classic method in machine learning to balance model complexity and performance. The improved attribute reduction method with rough set tries to get a trade-off between the number of features and empirical error, wherein the former is defined by rule confidence, and the latter is explored by the mutual information between the condition attribute subset and the decision attribute. In other words, model accuracy can be regarded as an empirical error term to measure the model's classification or prediction performance, and the model complexity as a penalty term to characterize the size of the reduction subset. To implement this approach, genetic algorithm is employed as a heuristic search technique, to obtain the optimal reduction subset. Several comparative experiments are performed on ten UCI datasets to evaluate the model's classification accuracy and the size of the reduction subset. The experimental results indicate the method proposed in this paper with better generalization ability compared to other traditional algorithms under the condition of equal reducted set length.