Goal Alignment: Re-analyzing Value Alignment Problems Using Human-Aware AI

Authors: Malek Mechergui, Sarath Sreedharan

AAAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental For evaluation, we ran our method on a set of problems selected from standard IPC benchmark problems (International Planning Competition 2011).
Researcher Affiliation Academia Malek Mechergui, Sarath Sreedharan Colorado State University {Malek.Mechergui, Sarath.Sreedharan}@colostate.edu
Pseudocode Yes Algorithm 1: An approximation-based algorithm to find a solution to a HAGL
Open Source Code Yes The code for the experiments can be found at: https://github.com/HAPILab/Goal Alignment.
Open Datasets Yes For evaluation, we ran our method on a set of problems selected from standard IPC benchmark problems (International Planning Competition 2011). International Planning Competition. 2011. IPC Competition Domains. https://goo.gl/i35bxc.
Dataset Splits No The paper discusses using "standard IPC benchmark problems" and creating "goal specification provided to the robot by randomly deleting a predicate from the goal specification" but does not provide specific dataset split information (e.g., train/validation/test percentages or counts) for reproduction.
Hardware Specification Yes All experiments were run on a linux Alma Linux 8.9 machine with 32GB ram and 16 Intel(R) Xeon(R) 2.60GHz CPUs.
Software Dependencies No The paper mentions using "Fast Downward planner" and "A-star search with LMcut heuristic" but does not provide specific version numbers for these software dependencies.
Experiment Setup Yes we set β to one for probability calculation.