Abstract
Deep neural networks (DNNs) can be useful within the marine robotics field, but their utility value is restricted by their black-box nature. Explainable artificial intelligence methods attempt to understand how such black-boxes make their decisions. In this work, linear model trees (LMTs) are used to approximate the DNN controlling an autonomous surface vessel (ASV) in a simulated environment and then run in parallel with the DNN to give explanations in the form of feature attributions in real-time. How well a model can be understood depends not only on the explanation itself, but also on how well it is presented and adapted to the receiver of said explanation. Different end-users may need both different types of explanations, as well as different representations of these. The main contributions of this work are (1) significantly improving both the accuracy and the build time of a greedy approach for building LMTs by introducing ordering of features in the splitting of the tree, (2) giving an overview of the characteristics of the seafarer/operator and the developer as two different end-users of the agent and receiver of the explanations, and (3) suggesting a visualization of the docking agent, the environment, and the feature attributions given by the LMT for when the developer is the end-user of the system, and another visualization for when the seafarer or operator is the end-user, based on their different characteristics.
Highlights
Machine learning is the sub-field of artificial intelligence (AI) dedicated to self-learning systems that use data to adjust their predictions
Among the most remarkable advancements in machine learning methods has been the evolution from artificial neural networks to deep architectures, known as deep neural networks (DNNs), forming the class of deep learning [1,2]. reinforcement learning (RL) is a branch of machine learning where an agent learns a strategy, referred to as a policy, which the agent uses to interact with an environment based on an evaluation of the agent’s interactions with the environment [3], called rewards
As our aim is for the linear model trees (LMTs) to be a faithful explanation model for the deep reinforcement learning (DRL) model, the loss in Equation (10) is calculated as the mean squared error (MSE) between the prediction of the DRL model and that of the linear function fitted by linear regression in the leaf nodes
Summary
Machine learning is the sub-field of artificial intelligence (AI) dedicated to self-learning systems that use data to adjust their predictions. DRL has shown to be a very useful tool for accomplishing difficult tasks in robotics, one advantage being that it does not require a mathematical model of the agent or the environment. In [17], a DRL-agent was trained to perform docking of an ASV in a simulated environment based on Trondheim harbor. Even though the agent showed promising and rather convincing results, its applicability to real-life problems is reduced by the lack of understanding of how the DNN-policy makes its decisions. This is because the many parameters and interconnections of DNNs make their inner workings hard for humans to interpret. Works in simulated environ- Works with the physical vessel, ments or digital twins without with risks for material damage risk of physical damage and potentially personnel injury
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
