There is a central dilemma embedded in the relationship between teachers and researchers. Teachers know the story of the classroom well, but they are seldom asked to tell their stories, nor do they usually have the opportunity. Researchers, on the other hand, are skilled at telling certain things about classrooms, but they often miss the central stories that are there. This divergence can lead to different opinions on what teaching is about and what is important within it. To bridge this gap, we describe an approach which puts the teacher and the student at the centre. With respect to emotional and behavioural problems of students, we underline the notion of student-teacher compatibility, deriving from theories emphasizing the transactional/reciprocal nature of human behaviour. One of the aims of the Lectorship and Knowledge Network Behavioural Problems in School Practice, is to identify at-risk-teachers (i.e. those most vulnerable to the presence of behaviourally challenging students and parents) so that interventions, both in initial teacher training as well as in inservice training can be applied to help them develop adequate attitudes and coping-skills. In clinical supervision, peer coaching or reflective practice, these teachers can be helped to consider in what way student and parental problem behaviour contribute to their loss of satisfaction, their feelings of self doubt, perceived disruption of the teaching process, and their frustration working with parents.
The number of applications in which industrial robots share their working environment with people is increasing. Robots appropriate for such applications are equipped with safety systems according to ISO/TS 15066:2016 and are often referred to as collaborative robots (cobots). Due to the nature of human-robot collaboration, the working environment of cobots is subjected to unforeseeable modifications caused by people. Vision systems are often used to increase the adaptability of cobots, but they usually require knowledge of the objects to be manipulated. The application of machine learning techniques can increase the flexibility by enabling the control system of a cobot to continuously learn and adapt to unexpected changes in the working environment. In this paper we address this issue by investigating the use of Reinforcement Learning (RL) to control a cobot to perform pick-and-place tasks. We present the implementation of a control system that can adapt to changes in position and enables a cobot to grasp objects which were not part of the training. Our proposed system uses deep Q-learning to process color and depth images and generates an (Formula presented.) -greedy policy to define robot actions. The Q-values are estimated using Convolution Neural Networks (CNNs) based on pre-trained models for feature extraction. To reduce training time, we implement a simulation environment to first train the RL agent, then we apply the resulting system on a real cobot. System performance is compared when using the pre-trained CNN models ResNext, DenseNet, MobileNet, and MNASNet. Simulation and experimental results validate the proposed approach and show that our system reaches a grasping success rate of 89.9% when manipulating a never-seen object operating with the pre-trained CNN model MobileNet.
Dementie is een wereldwijd gezondheidsprobleem, met enorme impact op de gezondheidszorg en economie. In Nederland is het aantal mensen met dementie de laatste jaren snel gegroeid en dit aantal zal door vergrijzing verder toenemen. De diagnose dementie is complex en vaak pas definitief na kostbaar en invasief onderzoek. Het introduceren van gemakkelijk uit te voeren tests in de eerstelijn kan bijdragen aan een verbeterde vroegtijdige herkenning en behandeling van dementie binnen de bredere bevolking. Een potentieel veelbelovende benadering is vroege detectie van retinale (netvlies) veranderingen in het oog met behulp van kunstmatige intelligentie (KI). Onderzoek toont aan dat neurale netwerken, een component van KI, subtiele afwijkingen in de retina kunnen detecteren die gerelateerd zijn aan dementie. Een algoritme, ontwikkeld door Cheung et al., toont veelbelovende resultaten op basis van retinale foto's. Het Geheugencentrum van het Jeroen Bosch Ziekenhuis heeft interesse in het gebruik van neurale netwerktechnologieën bij de diagnose van dementie. Het uitvoeren van een implementatieonderzoek gaat echter gepaard met uitdagingen op het gebied van dataverzameling, en daarnaast dienen zorgvuldige ethische overwegingen plaats te vinden. Om implementatieonderzoek in de toekomst mogelijk te maken, wil het Jeroen Bosch Ziekenhuis, samen met Biotactical BV en Avans Hogeschool verkennend onderzoek doen naar de nauwkeurigheid en praktische toepasbaarheid van het algoritme van Cheung et al., gebruikmakend van bestaande datasets. Daarnaast zal een juridisch en ethisch raamwerk worden ontworpen met richtlijnen voor een implementatieonderzoek van deze software in zorginstellingen. Het doel van dit verkennend onderzoeksproject is om vertrouwen op te bouwen bij belanghebbenden en 'lessons learned' op te nemen in een latere implementatiefase.
