2000 character limit reached
Guided Game Level Repair via Explainable AI
Published 30 Oct 2024 in cs.AI and cs.LG | (2410.23101v2)
Abstract: Procedurally generated levels created by machine learning models can be unsolvable without further editing. Various methods have been developed to automatically repair these levels by enforcing hard constraints during the post-processing step. However, as levels increase in size, these constraint-based repairs become increasingly slow. This paper proposes using explainability methods to identify specific regions of a level that contribute to its unsolvability. By assigning higher weights to these regions, constraint-based solvers can prioritize these problematic areas, enabling more efficient repairs. Our results, tested across three games, demonstrate that this approach can help to repair procedurally generated levels faster.
- Bradski, G. 2000. The OpenCV Library. Dr. Dobb’s Journal of Software Tools.
- Image-to-level: Generation and repair. In Proceedings of the AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment, volume 16, 189–195.
- Cooper, S. 2022a. Constraint-based 2d tile game blending in the sturgeon system. In Proceedings of the Experimental AI in Games Workshop.
- Cooper, S. 2022b. Sturgeon: tile-based procedural level generation via learned and designed constraints. Proceedings of the AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment, 18(1): 26–36.
- Literally Unplayable: On Constraint-Based Generation of Uncompletable Levels. In Proceedings of the 19th International Conference on the Foundations of Digital Games, 1–8.
- Pathfinding Agents for Platformer Level Repair. In AIIDE Workshops.
- Parallelizing the dual revised simplex method. Mathematical Programming Computation, 10(1): 119–142.
- PySAT: a Python toolkit for prototyping with SAT oracles. In Theory and Applications of Satisfiability Testing – SAT 2018, 428–437.
- RC2: an efficient MaxSAT solver. Journal on Satisfiability, Boolean Modeling and Computation, 11(1): 53–64.
- Autoencoders for level generation, repair, and recognition. In Proceedings of the ICCC workshop on computational creativity and games, volume 9.
- Kenney. 2022. Free game assets. https://www.kenney.nl/assets.
- Captum: A unified and generic model interpretability library for pytorch. arXiv preprint arXiv:2009.07896.
- A cardinality solver: more expressive constraints for free. In Cimatti, A.; and Sebastiani, R., eds., Theory and Applications of Satisfiability Testing – SAT 2012, Lecture Notes in Computer Science, 485–486. Berlin, Heidelberg: Springer. ISBN 978-3-642-31612-8.
- A Unified Approach to Interpreting Model Predictions. In Guyon, I.; Luxburg, U. V.; Bengio, S.; Wallach, H.; Fergus, R.; Vishwanathan, S.; and Garnett, R., eds., Advances in Neural Information Processing Systems, volume 30. Curran Associates, Inc.
- A Unified Approach to Interpreting Model Predictions. In Guyon, I.; Luxburg, U. V.; Bengio, S.; Wallach, H.; Fergus, R.; Vishwanathan, S.; and Garnett, R., eds., Advances in Neural Information Processing Systems 30, 4765–4774. Curran Associates, Inc.
- MSCG: robust core-guided MaxSAT solving. Journal on Satisfiability, Boolean Modeling and Computation, 9(1): 129–134.
- Controllable and coherent level generation: A two-pronged approach. In Experimental AI in games workshop.
- Neutronized. 2016. Super Cat Tales. Game [iPhone].
- Nintendo. 1985. Super Mario Bros. Game [NES].
- Rai, A. 2020. Explainable AI: From black box to glass box. Journal of the Academy of Marketing Science, 48: 137–141.
- Shapley, L. S.; et al. 1953. A value for n-person games.
- Not just a black box: Learning important features through propagating activation differences. arXiv preprint arXiv:1605.01713.
- A novel cnet-assisted evolutionary level repairer and its applications to Super Mario Bros. In 2020 IEEE Congress on Evolutionary Computation (CEC), 1–10. IEEE.
- Experiments in map generation using Markov chains. In International Conference on Foundations of Digital Games.
- Controllable Procedural Content Generation via Constrained Multi-Dimensional Markov Chain Sampling. In IJCAI, 780–786.
- Super mario as a string: Platformer level generation via lstms. arXiv preprint arXiv:1603.00930.
- Mcmcts pcg 4 smb: Monte carlo tree search to guide platformer level generation. In Proceedings of the AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment, volume 11, 68–74.
- Procedural content generation via machine learning (PCGML). IEEE Transactions on Games, 10(3): 257–270.
- The VGLC: The Video Game Level Corpus. arXiv:1606.07487 [cs].
- Axiomatic Attribution for Deep Networks. arXiv:1703.01365.
- The SciPy community. 2024. scipy.optimize.milp. https://docs.scipy.org/doc/scipy/reference/generated/scipy.optimize.milp.html.
- Search-Based Procedural Content Generation: A Taxonomy and Survey. IEEE Transactions on Computational Intelligence and AI in Games, 3(3): 172–186.
- SciPy 1.0: fundamental algorithms for scientific computing in Python. Nature methods, 17(3): 261–272.
- Evolving mario levels in the latent space of a deep convolutional generative adversarial network. In Proceedings of the genetic and evolutionary computation conference, 221–228.
- Video Game Level Repair via Mixed Integer Linear Programming. In Artificial Intelligence and Interactive Digital Entertainment Conference.
- 3.13 Explainable AI for Games. Human-Game AI Interaction, 73.
Paper Prompts
Sign up for free to create and run prompts on this paper using GPT-5.
Top Community Prompts
Collections
Sign up for free to add this paper to one or more collections.