There is emerging evidence that the performance of risk assessment instruments is weaker when used for clinical decision‐making than for research purposes. For instance, research has found lower agreement between evaluators when the risk assessments are conducted during routine practice. We examined the field interrater reliability of the Short‐Term Assessment of Risk and Treatability: Adolescent Version (START:AV). Clinicians in a Dutch secure youth care facility completed START:AV assessments as part of the treatment routine. Consistent with previous literature, interrater reliability of the items and total scores was lower than previously reported in non‐field studies. Nevertheless, moderate to good interrater reliability was found for final risk judgments on most adverse outcomes. Field studies provide insights into the actual performance of structured risk assessment in real‐world settings, exposing factors that affect reliability. This information is relevant for those who wish to implement structured risk assessment with a level of reliability that is defensible considering the high stakes.
DOCUMENT
Author-supplied abstract: Developing large-scale complex systems in student projects is not common, due to various constraints like available time, student team sizes, or maximal complexity. However, we succeeded to design a project that was of high complexity and comparable to real world projects. The execution of the project and the results were both successful in terms of quality, scope, and student/teacher satisfaction. In this experience report we describe how we combined a variety of principles and properties in the project design and how these have contributed to the success of the project. This might help other educators with setting up student projects of comparable complexity which are similar to real world projects.
DOCUMENT
In this post I give an overview of the theory, tools, frameworks and best practices I have found until now around the testing (and debugging) of machine learning applications. I will start by giving an overview of the specificities of testing machine learning applications.
LINK
TU Delft, in collaboration with Gravity Energy BV, has conducted a feasibility study on harvesting electric energy from wind and vibrations using a wobbling triboelectric nanogenerator (WTENG). Unlike conventional wind turbines, the WTENG converts wind/vibration energy into contact-separation events through a wobbling structure and unbalanced mass. Initial experimental findings demonstrated a peak power density of 1.6 W/m² under optimal conditions. Additionally, the harvester successfully charged a 3.7V lithium-ion battery with over 4.5 μA, illustrated in a self-powered light mast as a practical demonstration in collaboration with TimberLAB. This project aims to advance this research by developing a functioning prototype for public spaces, particularly lanterns, in partnership with TimberLAB and Gravity Energy. The study will explore the potential of triboelectric nanogenerators (TENG) and piezoelectric materials to optimize energy harvesting efficiency and power output. Specifically, the project will focus on improving the WTENG's output power for practical applications by optimizing parameters such as electrode dimensions and contact-separation quality. It will also explore cost-effective, commercially available materials and best fabrication/assembly strategies to simplify scalability for different length scales and power outputs. The research will proceed with the following steps: Design and Prototype Development: Create a prototype WTENG to evaluate energy harvesting efficiency and the quantity of energy harvested. A hybrid of TENG and piezoelectric materials will be designed and assessed. Optimization: Refine the system's design by considering the scaling effect and combinations of TENG-piezoelectric materials, focusing on maximizing energy efficiency (power output). This includes exploring size effects and optimal dimensions. Real-World Application Demonstration: Assess the optimized system's potential to power lanterns in close collaboration with TimberLAB, DVC Groep BV and Gravity Energy. Identify key parameters affecting the efficiency of WTENG technology and propose a roadmap for its exploitation in other applications such as public space lighting and charging.
Road freight transport contributes to 75% of the global logistics CO2 emissions. Various European initiatives are calling for a drastic cut-down of CO2 emissions in this sector [1]. This requires advanced and very expensive technological innovations; i.e. re-design of vehicle units, hybridization of powertrains and autonomous vehicle technology. One particular innovation that aims to solve this problem is multi-articulated vehicles (road-trains). They have a smaller footprint and better efficiency of transport than traditional transport vehicles like trucks. In line with the missions for Energy Transition and Sustainability [2], road-trains can have zero-emission powertrains leading to clean and sustainable urban mobility of people and goods. However, multiple articulations in a vehicle pose a problem of reversing the vehicle. Since it is extremely difficult to predict the sideways movement of the vehicle combination while reversing, no driver can master this process. This is also the problem faced by the drivers of TRENS Solar Train’s vehicle, which is a multi-articulated modular electric road vehicle. It can be used for transporting cargo as well as passengers in tight environments, making it suitable for operation in urban areas. This project aims to develop a reverse assist system to help drivers reverse multi-articulated vehicles like the TRENS Solar Train, enabling them to maneuver backward when the need arises in its operations, safely and predictably. This will subsequently provide multi-articulated vehicle users with a sustainable and economically viable option for the transport of cargo and passengers with unrestricted maneuverability resulting in better application and adding to the innovation in sustainable road transport.
A huge amount of data are being generated, collected, analysed and distributed in a fast pace in our daily life. This data growth requires efficient techniques for analysing and processing high volumes of data, for which preserving privacy effectively is a crucial challenge and even a key necessity, considering the recently coming into effect privacy laws (e.g., the EU General Data Protection Regulation-GDPR). Companies and organisations in their real-world applications need scalable and usable privacy preserving techniques to support them in protecting personal data. This research focuses on efficient and usable privacy preserving techniques in data processing. The research will be conducted in different directions: - Exploring state of the art techniques. - Designing and applying experiments on existing tool-sets. - Evaluating the results of the experiments based on the real-life case studies. - Improving the techniques and/or the tool to meet the requirements of the companies. The proposal will provide results for: - Education: like offering courses, lectures, students projects, solutions for privacy preservation challenges within the educational institutes. - Companies: like providing tool evaluation insights based on case studies and giving proposals for enhancing current challenges. - Research centre (i.e., Creating 010): like expanding its expertise on privacy protection technologies and publishing technical reports and papers. This research will be sustained by pursuing following up projects actively.
