No abstract

The segmentation of nanoscale electron microscopy (EM) images is crucial but challenging in connectomics. Recent advances in deep learning have demonstrated the significant potential of automatic segmentation for tera-scale EM images. However, none of the existing segmentation methods are error-free, and they require proofreading, which is typically implemented as an interactive, semi-automatic process via manual intervention. Herein, we propose a fully automatic proofreading method based on reinforcement learning. The main idea is to model the human decision process in proofreading using a reinforcement agent to achieve fully automatic proofreading. We systematically design the proposed system by combining multiple reinforcement learning agents in a hierarchical manner, where each agent focuses only on a specific task while preserving dependency between agents. Furthermore, we also demonstrate that the episodic task setting of reinforcement learning can efficiently manage a combination of merge and split errors concurrently presented in the input. We demonstrate the efficacy of the proposed system by comparing it with state-of-the-art proofreading methods using various testing examples.

The rising demand for mobility in the 21st century creates a challenge for interdisciplinary researchers. As a result, the number of papers devoted to the application of agent-based technologies in the transportation engineering domain has grown enormously. However, there is still a need for modelling platforms that are capable of exploring the influence of different psychological factors on individual decision-making. By utilising our current mobility simulator -BedDeM, we propose an experimental method to test and investigate the impact of core determinants in Triandis' Theory of Interpersonal Behaviour on the usage of different transportation modes. Comparing the results with a calibrated population of Swiss household data, we conclude that Intention and Affect have a positive effect on the usage of private vehicles, while Habit and Social factors can encourage people to travel with public or soft transportation modes.

No abstract

The development of efficient electric vehicle (EV) charging infrastructure requires modelling of consumer demand at an appropriate level of detail. Since only limited information about real customers is available, most simulations employ a stochastic approach by combining known or estimated business features (e.g. arrival and departure time, requested amount of energy) with random variations. However, these models in many cases do not include factors that deal with the social characteristics of EV users, while others do not emphasise on the economic elements. In this work, we introduced a more detailed demand model employing a modal choice simulation framework based on Triandis’ Theory of Interpersonal Behaviour, which can be calibrated by empirical data and is capable of combining a diverse number of determinants in human decision-making. By applying this model on Switzerland mobility domain, an analysis on three of the most popular EV incentives from both supply and demand sides is provided, which aims for a better understanding of electro-mobility systems by relating its causes and effects.