Ethically Compliant Sequential Decision Making

Authors: Justin Svegliato, Samer B. Nashed, Shlomo Zilberstein11657-11665

AAAI 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Finally, we demonstrate the accuracy and usability of our approach in a set of autonomous driving simulations and a user study of planning and robotics experts.
Researcher Affiliation Academia Justin Svegliato Samer B. Nashed Shlomo Zilberstein College of Information and Computer Sciences University of Massachusetts Amherst {jsvegliato,snashed,shlomo}@cs.umass.edu
Pseudocode No The paper describes the proposed approach and its components conceptually and mathematically but does not include any explicit pseudocode blocks or algorithms.
Open Source Code Yes Our open source library, Morality.js, available on the website https://www.moralityjs.com with the customizable grid world environment dashboard in Figure 3, was used in all experiments (Svegliato, Nashed, and Zilberstein 2020a,b).
Open Datasets No The paper describes a simulated city environment ('city in Figure 2') used for the autonomous driving simulations, but it does not provide a specific link, DOI, repository name, or formal citation for this environment as a publicly available or open dataset.
Dataset Splits No The paper describes experimental scenarios (autonomous driving simulations, user study) but does not provide specific details on how data was split into training, validation, and test sets. The 'data' in this context is the simulation environment and scenarios, not a traditional dataset with predefined splits.
Hardware Specification No The paper does not provide specific hardware details (e.g., CPU, GPU models, memory, or cloud instance types) used for running the experiments.
Software Dependencies No The paper mentions 'Our open source library, Morality.js' but does not provide specific version numbers for this or any other software dependencies like programming languages or libraries.
Experiment Setup Yes Each navigation task can use a different start location λ0 Λ and goal location λg Λ based on the city in Figure 2. The speed limits of city streets, county roads, and highways are 25, 45, and 75 MPH. The probability Pr(Θ = θ) of observing light or heavy pedestrian traffic θ Θ is 0.8 and 0.2. A low, normal, and high speed is 10 MPH under, at, and 10 MPH above the speed limit. Turning onto a road ω Ωfrom a location λ Λ requires 5 seconds. Accelerating 10 MPH requires 2 seconds. Cruising requires a time equal to the distance of the road ω Ωdivided by the speed σ Σ. Staying at a location λ Λ other than the goal location λg Λ requires 120 seconds. Each ethical framework can use different settings. For DCT, the forbidden states F can be just hazardous states H or both hazardous states H and inconsiderate states I. For PFD, the tolerance τ = ϵ can be the limit ϵ = 3, ϵ = 6, or ϵ = 9. For VE, the moral trajectories can be just cautious trajectories C or both cautious trajectories C and proactive trajectories P that avoid any highway road states and the set of populated location states.