Accenture (Switzerland)

This quality improvement project was designed to close a gap in capturing health literacy in adult patients with heart failure discharged from a hospital system. An opportunity existed to investigate if health literacy levels contribute to individuals' social factors and overall health. A priority focus was evident across healthcare organizations to address the social determinants of health (SDOH) for the populations served. Throughout the literature, limited health literacy is associated with higher hospital readmission rates. Pre-implementation data included 2023 heart failure readmission rate (19.02%) and baseline SDOH screenings (9,500) which noted housing, food insecurities, and transportation as greatest needs. There was no prior data analysis of how social factors may have impacted readmissions. Patients with recent hospitalization in Acute Care Hospitals with a known heart failure diagnosis; received health literacy screenings by the Registered Nurse (RN) Transitional Care Managers (TCMs) during their post-hospital discharge follow-up calls. Pre-implementation period began in August 2024. Project launched in November 2024 with RN TCMs conducting assessments using the BRIEF (Health Literacy Screening Tool) during post-hospital discharge calls. Additional follow-up calls were required as the intervention for patients with limited literacy during their 30-day post-hospitalization transition period. Data collection occurred from November 2024 through June 2025. The 8-month period included 1,543 high-risk hospital discharges, among them 580 patients with heart failure screened for health literacy; 517 patients had adequate health literacy, 44 had marginal health literacy, and 19 had limited health literacy. The intervention of additional follow-up calls for patients with limited health literacy achieved statistical significance in reducing readmissions. By implementing the BRIEF screening tool within an existing RN TCM workflow, it allowed for RN TCMs to interview patients with important questions during a vulnerable time for patients - transitioning from hospital to home. Expansion of health literacy assessments is vital for patients with any chronic condition, not just heart failure. RN TCMs identify the necessary interventions needed to support their patients and develop care plans. It is especially important to understand the needs specifically for patients with limited health literacy. Care management plays a pivotal role in patient advocacy by providing education and coordinating directly with providers to support patients' journey to better health.

Wisconsin’s construction industry plays a significant role in the state’s economy, supporting 380,000 jobs and contributing 15.8% to the state's GDP. The manufactured and modular building sector alone generates $2.65 billion annually and supports 26,000 jobs. Optimizing and managing modular construction processes reduces waste and enhances efficiency, making it key area for technological advancement. This paper proposes an integrated building information modeling (I-BIM) platform to enhance project management and promote sustainable construction. The I-BIM platform incorporates 3D visualization, 4D scheduling, 5D cost estimation, and 6D sustainability. It automates the generation of manufacturing drawings, reports, and an integrated database, enabling better project control, reducing rework, and optimizing energy use. The I-BIM model facilitates automation and integration across design and construction, using tools like Autodesk Revit, Navisworks, and INSIGHT-360. This platform aims to enhance waste reduction, project control, and stakeholder communication throughout the project lifecycle, supporting the broader goal of sustainable construction.

In today's world many students use a variety of digital resources when they are learning; however, these resources lack interoperatability. As a result of this many students have disjointed, and inefficient ways to study and complete their homework. This paper introduces a new student focused platform called StudyFlow that has been developed under ULEF to bring together the disparate pieces of the learning process into one cohesive, adaptable, and collaborative system for both teachers and students. The ULEF framework utilizes a cyclical process of setting goals, creating a plan, implementing a plan, reflecting on the implementation of a plan, and making any necessary adjustments to the plan. StudyFlow includes structured subject planning, adaptable daily schedules, active learning activities, an artificial intelligence based tutoring module for individualized support and on-going information about learner progress, and analytical scores for assessing strengths and weaknesses in learner performance. The modular architecture of the platform allows for the platform to be scalable across all subject areas, compliant with standards-based content specifications, and have secure data transfer between devices. The performance and progress of each learner will be measured over long periods of time utilizing metrics such as percentage of tasks completed, adherence to the plan, accuracy of solutions, and time management efficiency. The study will evaluate usability, learner interaction, and the effectiveness of the study process. The study will show that AI driven integrated workflows will enable students to more consistently learn, self-direct their learning, and achieve academically measurable outcomes.

This study aims at pixel-wise defect segmentation in Printed Circuit Boards (PCBs) using PCB Defects dataset using U-Net architecture. The primary objective is to demonstrate that lightweight segmentation architectures can effectively replace computationally expensive bounding-box models for resource-constrained environments. PCBs are critical components of electronic devices, and minor defects in these can lead to mal-functions. Automated diagnosis and rectification of PCB defects would need not only the classification and bounding-box-wise defect segmentation but also the pixel-wise region extraction from the image. While some studies exist on PCB defect segmentation, they heavily relied on bounded-box masks or resource-expensive frameworks such as YOLO. In this study, an efficient pixel-mask generation framework is proposed based on Hadamard-filtered difference imaging to generate pixel-level binary masks for PCB defects dataset without manual annotation. Then, a U-Net model is developed by training with this generated dataset. The evaluated metrics on the test set include 92% precision, 96% recall and 89% of IoU, which show competitive performance of the segmentation model. This study demonstrates that competing accuracies could be achieved for pixel-wise segmentation of PCB defects even with a simple U-Net model for resource-constrained devices.

Shear bands dictate the failure mechanisms of alloys across various strain rates and limit the damage tolerance of the alloy. While short-range amorphization has the potential to mitigate shear effects, it has thus far been confined to the nanoscale. Here, we extend amorphization to the micrometer scale, fundamentally replacing shear-dominated failure in multi-principal element alloy micropillars. We implement continuous compression strain-training from low to high strain rates, generating a top-down high-density dislocation gradient that drives the formation of a topological disorder network, extending over one-third of the micropillar height, which we define as hyper-range amorphization. Within the amorphous bands, atoms exhibit dynamic disorder, and the lattice rearranges and recovers, dissipating shear stress. The alloy achieves an ultimate compressive strength of ceramic level ( ~ 6.5 GPa), while maintaining ~59.1% plasticity. This work reveals a strain engineering-based mechanical mechanism for extending amorphization, establishing it as a viable pathway to enhancing the structural stability and energy dissipation capacity of alloys.

During the holiday season, retail supply chains are challenged with managing increased order volumes, lower order processing times, and customer expectations. Warehouse Control Systems (WCS) operational resiliency and performance during this peak period are essential to maintaining fulfillment velocity and accuracy. This paper presents a Holiday Readiness Scorecard approach designed to assess the readiness of WCS within Retail Distribution Center (DC) environments. The scorecard evaluates the following six readiness dimensions: System Performance and Capacity, Integration Stability, Hardware Reliability, Operational Readiness, Monitoring and Support, and Contingency Planning. By measuring readiness across these six domains, the scorecard provides a framework for correlating system readiness, identifying system vulnerabilities, prioritizing mitigations, and aligning cross-functional teams during the holiday season. The adoption of the scorecard approach enables operational readiness based on data correlation and provides insight into business continuity during the retail industry’s most critical season.

The deepening of commercial operation of the power system has put forward higher requirements for the accuracy, timeliness, and safety of electricity bill settlement, which is a key power engineering issue to ensure the economic benefits of power grid enterprises and the stable operation of the electricity market. The traditional manual review method is no longer able to cope with the constantly emerging risks and challenges. This article aims to address the key challenge of transitioning from “identification” to “closed-loop governance” of electricity bill settlement risks. Starting from the actual needs of power engineering, we propose an innovative governance framework that integrates multimodal artificial intelligence (AI) and process intelligent agents. This framework utilizes multimodal AI to identify anomalies from diverse data sources (text contracts, bills, electricity meter images), and employs software agents to automatically assign disposal tasks, track rectification processes, and verify risk mitigation, thus forming a complete governance loop. A practical case in a city in Zhejiang Province shows that this scheme can significantly improve the accuracy and efficiency of risk identification and disposal, achieve closed-loop management of risks throughout the entire process, and provide a practical and scalable intelligent solution for risk management in power engineering.

Robot grasping is a critical area of research in robotics, which aims to develop algorithms and strategies that enable robots towards reliable grasping and manipulation of objects in various environments. The ability to grasp and handle objects dexterously has wide-ranging applications in manufacturing, logistics, and healthcare. Recent advances in deep learning have paved the way for learning-based approaches toward autonomous grasping and intelligent manipulation in structured and unstructured environments. However, majority of the existing methods are designed for two-finger robotic grippers and cannot be extended towards application on three-finger robotic grippers. Three-finger grippers can perform a wide variety of grasps, and thereby require more nuanced grasp representation as compared to existing approaches. In this work, we propose a novel method towards representing underactuated three-finger grasps. A new grasp data set is also presented, specifically designed to train and evaluate three-finger robotic grasps using the proposed novel representation. A state-of-the-art deep learning architecture was trained on the dataset towards developing an end-to-end grasp planning pipeline. The proposed pipeline uses color and depth images as inputs to predict the ideal grasp pose for a three-finger robotic gripper to grasp the target object. The trained network was validated through simulation and hardware experiments, and the results were compared with an existing state-of-the-art grasp planning approach to highlight the potential of this new technique for grasp representation and planning.

With the proportion of renewable energy exceeding 35% (data of the National Energy Administration in 2024), the power system is facing the challenge of expanding the peak valley difference to 4.2:1. This paper proposes a "market technology policy" three drive collaborative framework, and constructs the following innovative schemes: 1) based on the intelligent response algorithm of deep reinforcement learning (DRL) and model predictive control (MPC), the DRL-MPC algorithm achieves the accuracy of 92.3% ± 1.2% for minute load forecasting; 2) The mixed integer programming energy storage optimization model with electricity price incentive can reduce the allocation cost by 18.6%; 3) The flexible polymerization technology of virtual power plant (VPP) has been verified by TenneT grid in Germany, and the peak load has been successfully reduced by 14.7%. Through the empirical analysis of the United States, Germany and China, this paper reveals the coupling mechanism of market signals and technical means, and verifies that the collaborative scheme can improve the new energy consumption rate to 91.2%, and reduce the carbon emission intensity by 23.8% compared with the 2020 benchmark. The research provides an economic and reliable technical path for the new power system, and helps to realize the grid elastic upgrading under the "double carbon" goal.

A key concern of automatic process discovery is providing insights into business process performance. Process analysts are specifically interested in waiting times and delays for identifying opportunities to speed up processes. Against this backdrop, it is surprising that current techniques for automatic process discovery generate directly-follows graphs and comparable process models without representing the time axis explicitly. This paper presents four layout strategies for automatically constructing process models that explicitly align with a time axis. We exemplify our approaches for directly-follows graphs. We evaluate their effectiveness by applying them to real-world event logs with varying complexities. Our specific focus is on their ability to handle the trade-off between high control-flow abstraction and high consistency of temporal activity order. Our results show that timeline-based layouts provide benefits in terms of an explicit representation of temporal distances. They face challenges for logs with many repeating and concurrent activities. • This work proposes four layout strategies to align a process model to a linear time axis. • Compared to traditional process models, timeline-based process models can more effectively convey the temporal perspective of the model both in its entirety and local fragments, supporting the finding of possible bottlenecks. • Timeline-based process models face a challenge in balancing the trade-off between keeping a high model abstraction and a consistent model-timeline alignment.