Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
124 tokens/sec
GPT-4o
8 tokens/sec
Gemini 2.5 Pro Pro
47 tokens/sec
o3 Pro
5 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

A Survey on Large Language Model-Based Game Agents (2404.02039v2)

Published 2 Apr 2024 in cs.AI

Abstract: The development of game agents holds a critical role in advancing towards Artificial General Intelligence. The progress of LLMs offers an unprecedented opportunity to evolve and empower game agents with human-like decision-making capabilities in complex computer game environments. This paper provides a comprehensive overview of LLM-based game agents from a holistic viewpoint. First, we introduce the conceptual architecture of LLM-based game agents, centered around three core functional components: memory, reasoning and in/output. Second, we survey existing representative LLM-based game agents documented in the literature with respect to methodologies and adaptation agility across six genres of games, including adventure, communication, competition, cooperation, simulation, and crafting & exploration games. Finally, we present an outlook of future research and development directions in this burgeoning field. A curated list of relevant papers is maintained and made accessible at: https://github.com/git-disl/awesome-LLM-game-agent-papers.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (169)
  1. The development of embodied cognition: Six lessons from babies. Artificial life, 11(1-2):13–29, 2005.
  2. Training language models to follow instructions with human feedback. Advances in Neural Information Processing Systems, 35:27730–27744, 2022.
  3. Gpt-4 technical report. arXiv preprint arXiv:2303.08774, 2023.
  4. Gemini: a family of highly capable multimodal models. arXiv preprint arXiv:2312.11805, 2023.
  5. Cognitive architectures for language agents. arXiv preprint arXiv:2309.02427, 2023.
  6. Language modeling is compression. arXiv preprint arXiv:2309.10668, 2023.
  7. Artificial general intelligence, vol. 2. Springer, 2007.
  8. Goertzel, B. Artificial general intelligence: concept, state of the art, and future prospects. Journal of Artificial General Intelligence, 5(1):1, 2014.
  9. IBM. Deep blue. https://www.ibm.com/history/deep-blue.
  10. Mastering the game of go with deep neural networks and tree search. nature, 529(7587):484–489, 2016.
  11. Bayes’ bluff: Opponent modelling in poker. arXiv preprint arXiv:1207.1411, 2012.
  12. The arcade learning environment: An evaluation platform for general agents. Journal of Artificial Intelligence Research, 47:253–279, 2013.
  13. Starcraft ii: A new challenge for reinforcement learning. arXiv preprint arXiv:1708.04782, 2017.
  14. Mojang Studios. Minecraft. https://www.minecraft.net/en-us.
  15. Dota 2. https://www.dota2.com/home.
  16. Mastering the game of go without human knowledge. nature, 550(7676):354–359, 2017.
  17. OpenAI. Openai five. https://openai.com/research/openai-five, 2018. Accessed on: yyyy-mm-dd.
  18. DeepMind. Alphastar: Mastering the real-time strategy game starcraft ii. https://deepmind.google/discover/blog/alphastar-mastering-the-real-time-strategy-game-starcraft-ii, 2019.
  19. A survey of large language models. arXiv preprint arXiv:2303.18223, 2023.
  20. A survey on evaluation of large language models. ACM Transactions on Intelligent Systems and Technology, 2023.
  21. Challenges and applications of large language models. arXiv preprint arXiv:2307.10169, 2023.
  22. The rise and potential of large language model based agents: A survey. arXiv preprint arXiv:2309.07864, 2023.
  23. A survey on large language model based autonomous agents. arXiv preprint arXiv:2308.11432, 2023.
  24. Large language models empowered agent-based modeling and simulation: A survey and perspectives. arXiv preprint arXiv:2312.11970, 2023.
  25. Large language models and games: A survey and roadmap. arXiv preprint arXiv:2402.18659, 2024.
  26. Sweetser, P. Large language models and video games: A preliminary scoping review. arXiv preprint arXiv:2403.02613, 2024.
  27. Infocom. Zork I. http://ifdb.tads.org/viewgame?id=0dbnusxunq7fw5ro, 1980.
  28. Exploring large language models for communication games: An empirical study on werewolf. arXiv preprint arXiv:2309.04658, 2023.
  29. Large language models play starcraft ii: Benchmarks and a chain of summarization approach. arXiv preprint arXiv:2312.11865, 2023.
  30. Pokéllmon: A human-parity agent for pokémon battles with large language models, 2024.
  31. Cooperative open-ended learning framework for zero-shot coordination. arXiv preprint arXiv:2302.04831, 2023.
  32. Freeciv-web contributors. Freeciv-web. https://github.com/freeciv/freeciv-web, 2023.
  33. PrismarineJS. Mineflayer: Create minecraft bots with a powerful, stable, and high level javascript api. https://github.com/PrismarineJS/mineflayer, 2013.
  34. Towards general computer control: A multimodal agent for red dead redemption ii as a case study. arXiv preprint arXiv:2403.03186, 2024.
  35. Building cooperative embodied agents modularly with large language models. arXiv preprint arXiv:2307.02485, 2023.
  36. Llm-planner: Few-shot grounded planning for embodied agents with large language models. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 2998–3009. 2023.
  37. Learning transferable visual models from natural language supervision. In International conference on machine learning, pages 8748–8763. PMLR, 2021.
  38. Describe, explain, plan and select: Interactive planning with large language models enables open-world multi-task agents. arXiv preprint arXiv:2302.01560, 2023.
  39. —. Jarvis-1: Open-world multi-task agents with memory-augmented multimodal language models. arXiv preprint arXiv:2311.05997, 2023.
  40. Guiding pretraining in reinforcement learning with large language models. In International Conference on Machine Learning, pages 8657–8677. PMLR, 2023.
  41. Minedojo: Building open-ended embodied agents with internet-scale knowledge. Advances in Neural Information Processing Systems, 35:18343–18362, 2022.
  42. Clipcap: Clip prefix for image captioning. arXiv preprint arXiv:2111.09734, 2021.
  43. de Wynter, A. Will gpt-4 run doom? arXiv preprint arXiv:2403.05468, 2024.
  44. Creative agents: Empowering agents with imagination for creative tasks. arXiv preprint arXiv:2312.02519, 2023.
  45. Octopus: Embodied vision-language programmer from environmental feedback. arXiv preprint arXiv:2310.08588, 2023.
  46. A generalist agent. arXiv preprint arXiv:2205.06175, 2022.
  47. Steve-eye: Equipping llm-based embodied agents with visual perception in open worlds. In The Twelfth International Conference on Learning Representations. 2023.
  48. Infocom. Zork III. http://ifdb.tads.org/viewgame?id=vrsot1zgy1wfcdru, 1982.
  49. Scienceworld: Is your agent smarter than a 5th grader? arXiv preprint arXiv:2203.07540, 2022.
  50. Alfworld: Aligning text and embodied environments for interactive learning. arXiv preprint arXiv:2010.03768, 2020.
  51. Human-level play in the game of diplomacy by combining language models with strategic reasoning. Science, 378(6624):1067–1074, 2022.
  52. Avalon’s game of thoughts: Battle against deception through recursive contemplation. arXiv preprint arXiv:2310.01320, 2023.
  53. Pokergpt: An end-to-end lightweight solver for multi-player texas hold’em via large language model. arXiv preprint arXiv:2401.06781, 2024.
  54. Suspicion-agent: Playing imperfect information games with theory of mind aware gpt-4. arXiv preprint arXiv:2309.17277, 2023.
  55. Chessgpt: Bridging policy learning and language modeling. Advances in Neural Information Processing Systems, 36, 2024.
  56. Civrealm: A learning and reasoning odyssey in civilization for decision-making agents. arXiv preprint arXiv:2401.10568, 2024.
  57. Human memory: A proposed system and its control processes. In Psychology of learning and motivation, vol. 2, pages 89–195. Elsevier, 1968.
  58. Extending cognitive architecture with episodic memory. In AAAI, pages 1560–1564. 2007.
  59. Generative agents: Interactive simulacra of human behavior. In Proceedings of the 36th Annual ACM Symposium on User Interface Software and Technology, pages 1–22. 2023.
  60. Multi-stage episodic control for strategic exploration in text games. arXiv preprint arXiv:2201.01251, 2022.
  61. Agent-pro: Learning to evolve via policy-level reflection and optimization. arXiv preprint arXiv:2402.17574, 2024.
  62. Toward integrating cognitive linguistics and cognitive language processing. In Proceedings of the 14th International Conference on Cognitive Modeling (ICCM). 2016.
  63. Reflexion: an autonomous agent with dynamic memory and self-reflection. arXiv preprint arXiv:2303.11366, 2023.
  64. Working, declarative and procedural memory in specific language impairment. cortex, 48(9):1138–1154, 2012.
  65. Voyager: An open-ended embodied agent with large language models. arXiv preprint arXiv:2305.16291, 2023.
  66. Ghost in the minecraft: Generally capable agents for open-world enviroments via large language models with text-based knowledge and memory. arXiv preprint arXiv:2305.17144, 2023.
  67. Hierarchical auto-organizing system for open-ended multi-agent navigation. arXiv preprint arXiv:2403.08282, 2024.
  68. Non-player character decision-making in computer games. Artificial Intelligence Review, 56(12):14159–14191, 2023.
  69. Calypso: Llms as dungeon master’s assistants. In Proceedings of the AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment, vol. 19, pages 380–390. 2023.
  70. Rolellm: Benchmarking, eliciting, and enhancing role-playing abilities of large language models. arXiv preprint arXiv:2310.00746, 2023.
  71. Personallm: Investigating the ability of gpt-3.5 to express personality traits and gender differences. arXiv preprint arXiv:2305.02547, 2023.
  72. Chatharuhi: Reviving anime character in reality via large language model. arXiv preprint arXiv:2308.09597, 2023.
  73. Lamp: When large language models meet personalization. arXiv preprint arXiv:2304.11406, 2023.
  74. Camel: Communicative agents for" mind" exploration of large language model society. Advances in Neural Information Processing Systems, 36, 2024.
  75. An appraisal-based chain-of-emotion architecture for affective language model game agents. arXiv preprint arXiv:2309.05076, 2023.
  76. Emotionally numb or empathetic? evaluating how llms feel using emotionbench. arXiv preprint arXiv:2308.03656, 2023.
  77. Charactereval: A chinese benchmark for role-playing conversational agent evaluation. arXiv preprint arXiv:2401.01275, 2024.
  78. Character-llm: A trainable agent for role-playing. arXiv preprint arXiv:2310.10158, 2023.
  79. Towards reasoning in large language models: A survey. arXiv preprint arXiv:2212.10403, 2022.
  80. React: Synergizing reasoning and acting in language models. arXiv preprint arXiv:2210.03629, 2022.
  81. Chain-of-thought prompting elicits reasoning in large language models. Advances in Neural Information Processing Systems, 35:24824–24837, 2022.
  82. Large language models are zero-shot reasoners. Advances in neural information processing systems, 35:22199–22213, 2022.
  83. Tree of thoughts: Deliberate problem solving with large language models. arXiv preprint arXiv:2305.10601, 2023.
  84. Self-consistency improves chain of thought reasoning in language models. arXiv preprint arXiv:2203.11171, 2022.
  85. Graph of thoughts: Solving elaborate problems with large language models. arXiv preprint arXiv:2308.09687, 2023.
  86. Self-refine: Iterative refinement with self-feedback. Advances in Neural Information Processing Systems, 36, 2024.
  87. Proagent: Building proactive cooperative ai with large language models. arXiv preprint arXiv:2308.11339, 2023.
  88. Language models can solve computer tasks. Advances in Neural Information Processing Systems, 36, 2024.
  89. Large language model-powered smart contract vulnerability detection: New perspectives. arXiv preprint arXiv:2310.01152, 2023.
  90. Theory of mind. Current biology, 15(17):R644–R645, 2005.
  91. Kosinski, M. Theory of mind may have spontaneously emerged in large language models. arXiv preprint arXiv:2302.02083, 2023.
  92. Evaluating multi-agent coordination abilities in large language models. arXiv preprint arXiv:2310.03903, 2023.
  93. Omni: Open-endedness via models of human notions of interestingness. arXiv preprint arXiv:2306.01711, 2023.
  94. Spring: Studying papers and reasoning to play games. Advances in Neural Information Processing Systems, 36, 2024.
  95. Adarefiner: Refining decisions of language models with adaptive feedback, 2023.
  96. Adapt: As-needed decomposition and planning with language models. arXiv preprint arXiv:2311.05772, 2023.
  97. Swiftsage: A generative agent with fast and slow thinking for complex interactive tasks. Advances in Neural Information Processing Systems, 36, 2024.
  98. Language models as zero-shot planners: Extracting actionable knowledge for embodied agents. In International Conference on Machine Learning, pages 9118–9147. PMLR, 2022.
  99. S-agents: self-organizing agents in open-ended environment. arXiv preprint arXiv:2402.04578, 2024.
  100. Mindagent: Emergent gaming interaction. arXiv preprint arXiv:2309.09971, 2023.
  101. Plan4mc: Skill reinforcement learning and planning for open-world minecraft tasks. arXiv preprint arXiv:2303.16563, 2023.
  102. On the utility of learning about humans for human-ai coordination. Advances in neural information processing systems, 32, 2019.
  103. Microsoft Research. First textworld problems: The competition using text-based games to advance capabilities of ai agents, 2019.
  104. Interactive fiction games: A colossal adventure. In Proceedings of the AAAI Conference on Artificial Intelligence, vol. 34, pages 7903–7910. 2020.
  105. Llm-powered hierarchical language agent for real-time human-ai coordination. arXiv preprint arXiv:2312.15224, 2023.
  106. Rl-gpt: Integrating reinforcement learning and code-as-policy. arXiv preprint arXiv:2402.19299, 2024.
  107. Virtualhome: Simulating household activities via programs. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 8494–8502. 2018.
  108. Do as i can, not as i say: Grounding language in robotic affordances. arXiv preprint arXiv:2204.01691, 2022.
  109. War and peace (waragent): Large language model-based multi-agent simulation of world wars. arXiv preprint arXiv:2311.17227, 2023.
  110. Language models meet world models: Embodied experiences enhance language models. Advances in neural information processing systems, 36, 2024.
  111. Llama rider: Spurring large language models to explore the open world. arXiv preprint arXiv:2310.08922, 2023.
  112. Pre-trained language models for interactive decision-making. Advances in Neural Information Processing Systems, 35:31199–31212, 2022.
  113. Grounding large language models in interactive environments with online reinforcement learning. arXiv preprint arXiv:2302.02662, 2023.
  114. Proximal policy optimization algorithms. arXiv preprint arXiv:1707.06347, 2017.
  115. Enhance reasoning for large language models in the game werewolf. arXiv preprint arXiv:2402.02330, 2024.
  116. Reward design with language models. arXiv preprint arXiv:2303.00001, 2023.
  117. Motif: Intrinsic motivation from artificial intelligence feedback. arXiv preprint arXiv:2310.00166, 2023.
  118. Auto mc-reward: Automated dense reward design with large language models for minecraft. arXiv preprint arXiv:2312.09238, 2023.
  119. Text2reward: Automated dense reward function generation for reinforcement learning. arXiv preprint arXiv:2309.11489, 2023.
  120. Eureka: Human-level reward design via coding large language models. arXiv preprint arXiv:2310.12931, 2023.
  121. Lyfe agents: Generative agents for low-cost real-time social interactions. arXiv preprint arXiv:2310.02172, 2023.
  122. Hafner, D. Benchmarking the spectrum of agent capabilities. arXiv preprint arXiv:2109.06780, 2021.
  123. Keep calm and explore: Language models for action generation in text-based games. arXiv preprint arXiv:2010.02903, 2020.
  124. Can large language models play text games well? current state-of-the-art and open questions. arXiv preprint arXiv:2304.02868, 2023.
  125. Language agents with reinforcement learning for strategic play in the werewolf game. arXiv preprint arXiv:2310.18940, 2023.
  126. Avalonbench: Evaluating llms playing the game of avalon. arXiv e-prints, pages arXiv–2310, 2023.
  127. Cooperation on the fly: Exploring language agents for ad hoc teamwork in the avalon game. arXiv preprint arXiv:2312.17515, 2023.
  128. What if llms have different world views: Simulating alien civilizations with llm-based agents. arXiv preprint arXiv:2402.13184, 2024.
  129. Leveraging word guessing games to assess the intelligence of large language models. arXiv preprint arXiv:2310.20499, 2023.
  130. Gameeval: Evaluating llms on conversational games. arXiv preprint arXiv:2308.10032, 2023.
  131. Swarmbrain: Embodied agent for real-time strategy game starcraft ii via large language models. arXiv preprint arXiv:2401.17749, 2024.
  132. Emergent world representations: Exploring a sequence model trained on a synthetic task. arXiv preprint arXiv:2210.13382, 2022.
  133. Gupta, A. Are chatgpt and gpt-4 good poker players?–a pre-flop analysis. arXiv preprint arXiv:2308.12466, 2023.
  134. Humanoid agents: Platform for simulating human-like generative agents. arXiv preprint arXiv:2310.05418, 2023.
  135. Agentsims: An open-source sandbox for large language model evaluation. arXiv preprint arXiv:2308.04026, 2023.
  136. Regal: Refactoring programs to discover generalizable abstractions. arXiv preprint arXiv:2401.16467, 2024.
  137. Textworld: A learning environment for text-based games. In Computer Games: 7th Workshop, CGW 2018, Held in Conjunction with the 27th International Conference on Artificial Intelligence, IJCAI 2018, Stockholm, Sweden, July 13, 2018, Revised Selected Papers 7, pages 41–75. Springer, 2019.
  138. Counting to explore and generalize in text-based games. arXiv preprint arXiv:1806.11525, 2018.
  139. BBC. The hitchhiker’s guide to the galaxy text adventure: 30th anniversary edition. https://www.bbc.co.uk/programmes/articles/1g84m0sXpnNCv84GpN2PLZG/the-game-30th-anniversary-edition.
  140. Alfred: A benchmark for interpreting grounded instructions for everyday tasks. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 10740–10749. 2020.
  141. Babyai: A platform to study the sample efficiency of grounded language learning. arXiv preprint arXiv:1810.08272, 2018.
  142. Pre-trained language models as prior knowledge for playing text-based games. arXiv preprint arXiv:2107.08408, 2021.
  143. Language model-in-the-loop: Data optimal approach to learn-to-recommend actions in text games. arXiv preprint arXiv:2311.07687, 2023.
  144. Self-imitation learning for action generation in text-based games. In Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics, pages 703–726. 2023.
  145. Graph constrained reinforcement learning for natural language action spaces. arXiv preprint arXiv:2001.08837, 2020.
  146. Interactive fiction game playing as multi-paragraph reading comprehension with reinforcement learning. arXiv preprint arXiv:2010.02386, 2020.
  147. Bart: Denoising sequence-to-sequence pre-training for natural language generation, translation, and comprehension. arXiv preprint arXiv:1910.13461, 2019.
  148. Deciphering digital detectives: Understanding llm behaviors and capabilities in multi-agent mystery games. arXiv preprint arXiv:2312.00746, 2023.
  149. Wikipedia. Portable game notation. https://en.wikipedia.org/wiki/Portable_Game_Notation, 2023.
  150. Chess as a testbed for language model state tracking. In Proceedings of the AAAI Conference on Artificial Intelligence, vol. 36, pages 11385–11393. 2022.
  151. Opt: Open pre-trained transformer language models. arXiv preprint arXiv:2205.01068, 2022.
  152. Alphaholdem: High-performance artificial intelligence for heads-up no-limit poker via end-to-end reinforcement learning. In Proceedings of the AAAI Conference on Artificial Intelligence, vol. 36, pages 4689–4697. 2022.
  153. Too many cooks: Bayesian inference for coordinating multi-agent collaboration. Topics in Cognitive Science, 13(2):414–432, 2021.
  154. Watch-and-help: A challenge for social perception and human-ai collaboration. arXiv preprint arXiv:2010.09890, 2020.
  155. The threedworld transport challenge: A visually guided task-and-motion planning benchmark for physically realistic embodied ai. arXiv preprint arXiv:2103.14025, 2021.
  156. Threedworld: A platform for interactive multi-modal physical simulation. arXiv preprint arXiv:2007.04954, 2020.
  157. Maximum entropy population-based training for zero-shot human-ai coordination. In Proceedings of the AAAI Conference on Artificial Intelligence, vol. 37, pages 6145–6153. 2023.
  158. Significant Gravitas. AutoGPT.
  159. Ai2-thor: An interactive 3d environment for visual ai. arXiv preprint arXiv:1712.05474, 2017.
  160. Interactive gibson benchmark: A benchmark for interactive navigation in cluttered environments. IEEE Robotics and Automation Letters, 5(2):713–720, 2020.
  161. Habitat: A platform for embodied ai research. In Proceedings of the IEEE/CVF international conference on computer vision, pages 9339–9347. 2019.
  162. Behavior: Benchmark for everyday household activities in virtual, interactive, and ecological environments. In Conference on robot learning, pages 477–490. PMLR, 2022.
  163. Behavior-1k: A benchmark for embodied ai with 1,000 everyday activities and realistic simulation. In Conference on Robot Learning, pages 80–93. PMLR, 2023.
  164. Automatic goal generation for reinforcement learning agents. In International conference on machine learning, pages 1515–1528. PMLR, 2018.
  165. Self-instruct: Aligning language model with self generated instructions. arXiv preprint arXiv:2212.10560, 2022.
  166. Toolkengpt: Augmenting frozen language models with massive tools via tool embeddings. Advances in neural information processing systems, 36, 2024.
  167. Climbing towards nlu: On meaning, form, and understanding in the age of data. In Proceedings of the 58th annual meeting of the association for computational linguistics, pages 5185–5198. 2020.
  168. On the opportunities and risks of foundation models. arXiv preprint arXiv:2108.07258, 2021.
  169. Metaagents: Simulating interactions of human behaviors for llm-based task-oriented coordination via collaborative generative agents. arXiv preprint arXiv:2310.06500, 2023.
Citations (36)
View on Semantic Scholar

Summary

  • The paper presents a comprehensive framework that categorizes LLM-based game agents into six key components mirroring human cognition in gaming.
  • It evaluates agent performance across genres like adventure, competition, and simulation using varied metrics including task success rates and win rates.
  • The study outlines future directions by emphasizing advanced grounding in game environments to boost agent realism and further AGI progress.

Advances and Prospects in LLM-based Game Agents

Introduction to LLM-based Game Agents

The role of game agents in the advancement of AGI is undeniably crucial, with the rapid evolution of LLMs and their multimodal counterparts offering new pathways to enhance these agents' capabilities. This paper meticulously surveys the progress of LLM-based game agents (LLMGAs), defining their architecture and exploring their application across various game genres. By organizing the discussion into six core components—perception, memory, thinking, role-playing, action, and learning—the paper lays down a comprehensive framework for current methodologies and future research directions.

Core Components of LLMGAs

LLMGAs are distinguished by six functional components designed to mimic human cognitive processes:

  • Perception: LLMGAs leverage text and visual input to perceive game environments, employing methods like state variable access, external visual encoders, and multimodal LLMs for accurate perception.
  • Memory: Serving as an external storage, memory in LLMGAs holds past experiences and knowledge, crucial for strategy formulation.
  • Thinking: Incorporating reasoning and planning, this component analyses information, infers explanations, and strategizes actions based on game states.
  • Role-playing: By simulating specific roles, LLMGAs can generate believable behaviors and dialogues aligned with their characters within games.
  • Action: This translates decisions into executable game actions, with different strategies for manipulating game elements effectively.
  • Learning: LLMGAs improve over time through interactions with the game environment, employing methods like in-context feedback learning, supervised fine-tuning, and reinforcement learning.

Categorization of Games

The paper organizes games into six categories, assessing LLMGAs' performance and strategies within each:

  • Adventure Games: Exploring narrative-driven environments, where agents address complex quests through story interaction.
  • Communication Games: Focused on negotiation and deception to infer and conceal intentions among players.
  • Competition Games: Highlighting skill or strategy to outperform opponents, requiring advanced reasoning.
  • Cooperation Games: Where teamwork and collaborative problem-solving are key to achieving common goals.
  • Simulation Games: Simulating real-world scenarios, demanding realistic interaction and decision-making.
  • Crafting and Exploration Games: Encouraging creativity in resource gathering and item crafting within expansive environments.

Evaluation Metrics

Evaluating LLMGAs involves varied metrics across game types, including task success rate for adventure and simulation games, win rates for competition and communication games, and human evaluation for simulations of social interactions.

Implications and Future Directions

The advancement of LLMGAs reflects significant progress toward AGI, particularly in understanding and interacting with complex environments. Future efforts could focus on grounding LLMs in game environments for deeper contextual understanding, enabling knowledge discovery beyond current gameplay to unearth fundamental principles through experience, and enhancing agent society simulations for a closer representation of human social interactions.

Conclusion

This comprehensive examination of LLM-based game agents underscores their pivotal role in pushing the boundaries of AI research. By detailing the architecture, application, and evaluation of LLMGAs across a broad spectrum of game genres, the paper sets the stage for future advances that promise to bring us closer to realizing human-like AGI. The curated list of papers and ongoing updates serve as a valuable resource for researchers in exploring the burgeoning field of game-based AI research.

File Document Download Save Streamline Icon: https://streamlinehq.com PDF File Document Download Save Streamline Icon: https://streamlinehq.com Markdown
Github Logo Streamline Icon: https://streamlinehq.com

GitHub

  1. GitHub - git-disl/awesome-LLM-game-agent-papers: awesome LLM-based GAME agent papers (103 stars)
Youtube Logo Streamline Icon: https://streamlinehq.com

YouTube