no code implementations • 20 Feb 2024 • Sindre Benjamin Remman, Anastasios M. Lekkas
In this paper, we introduce a new approach for investigating the behavior modes of DRL policies, which involves utilizing dimensionality reduction and trajectory clustering in the latent space of neural networks.
no code implementations • 1 Mar 2022 • Vilde B. Gjærum, Ella-Lovise H. Rørvik, Anastasios M. Lekkas
The two main benefits of the proposed approach are: a) LMTs are transparent which makes it possible to associate directly the outputs (control actions, in our case) with specific values of the input features, b) LMTs are computationally efficient and can provide information in real-time.
no code implementations • 1 Mar 2022 • Vilde B. Gjærum, Inga Strümke, Ole Andreas Alsos, Anastasios M. Lekkas
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.
Explainable artificial intelligence Reinforcement Learning (RL)
no code implementations • 4 Nov 2021 • Sindre Benjamin Remman, Inga Strümke, Anastasios M. Lekkas
This partial causal ordering defines the causal relations between the features, and we specify this using domain knowledge about the lever control task.
Explainable artificial intelligence Reinforcement Learning (RL)
no code implementations • 7 Oct 2021 • Sindre Benjamin Remman, Anastasios M. Lekkas
To explain the decisions of the policy we use the SHAP method to create an explanation model based on the episodes done in the real-world environment.
no code implementations • 16 Jun 2021 • WenQi Cai, Arash B. Kordabad, Hossein N. Esfahani, Anastasios M. Lekkas, Sebastien Gros
In this work, we propose a Model Predictive Control (MPC)-based Reinforcement Learning (RL) method for Autonomous Surface Vehicles (ASVs).
no code implementations • 22 Mar 2021 • Arash Bahari Kordabad, Hossein Nejatbakhsh Esfahani, Anastasios M. Lekkas, Sébastien Gros
A scenario-tree robust MPC is used to handle potential failures of the ship thrusters.