Angry Birds as a Challenge for Artificial Intelligence

Authors: Jochen Renz, XiaoYu Ge, Rohan Verma, Peng Zhang

AAAI 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental The Angry Birds AI Competition has been held annually since 2012... We also summarise some highlights of past competitions, including a new competition track we introduced recently. After each AI competition, we hold a Man vs Machine Challenge to test if AI agents are already better than humans. In previous competitions, humans always won with a wide, but shrinking margin. In 2013, half of human participants were better than the best AI, while in 2014 it was a third.
Researcher Affiliation Academia Jochen Renz Xiao Yu Ge, Rohan Verma, Peng Zhang Research School of Computer Science The Australian National University jochen.renz@anu.edu.au
Pseudocode No The paper does not contain any pseudocode or clearly labeled algorithm blocks.
Open Source Code No The paper mentions that 'Starting in 2014, the best teams have made the source code of their agents available' for the competition, but it does not provide open-source code for any methodology presented by the authors of this paper.
Open Datasets No The paper refers to 'Angry Birds levels' as the basis for the competition, but it does not provide concrete access information (link, DOI, formal citation) for a publicly available or open dataset of these levels for research purposes.
Dataset Splits No The paper describes a competition involving 'new Angry Birds levels' but does not specify any training, validation, or test dataset splits.
Hardware Specification No The paper does not provide specific details about the hardware used for running experiments.
Software Dependencies No The paper mentions 'Box2D' as a game internal physics engine, but it does not provide specific ancillary software details with version numbers for any experimental setup.
Experiment Setup No The paper describes the competition and various AI approaches, but it does not provide specific experimental setup details such as hyperparameter values or training configurations.