Awesome AI Agents › Generative Agents Simulation
Voyager is an innovative open-ended embodied agent powered by large language models, specifically designed to operate within the Minecraft environment. It represents a significant advancement in AI by enabling lifelong learning and autonomous exploration without human intervention. The agent continuously explores the Minecraft world, acquiring diverse skills and making novel discoveries through an automatic curriculum that maximizes exploration. A key feature of Voyager is its ever-growing skill library, which stores executable code representing complex behaviors. This library allows the agent to retrieve and compose skills, enhancing its capabilities over time while mitigating issues like catastrophic forgetting. Voyager employs a novel iterative prompting mechanism that integrates environment feedback, execution errors, and self-verification to improve its programs. It interacts with GPT-4 through blackbox queries, eliminating the need for fine-tuning the model parameters. This approach results in temporally extended, interpretable, and compositional skills that compound the agent's abilities rapidly. Empirical results demonstrate Voyager's strong in-context lifelong learning capabilities and exceptional proficiency in Minecraft, outperforming prior state-of-the-art methods by significant margins in terms of unique items obtained, distances traveled, and speed in unlocking key milestones. The project provides a comprehensive codebase under the MIT License, supporting installation on multiple operating systems and requiring dependencies such as Python, Node.js, Minecraft, and Fabric mods. Users can run Voyager with an OpenAI GPT-4 API key and have options to resume learning from checkpoints or run specific tasks using a learned skill library. The repository includes detailed instructions for setup, usage, and task decomposition, making it accessible for research and experimentation. Voyager is a cutting-edge research project that pushes the boundaries of embodied AI agents by combining large language models with interactive environments, enabling continuous skill acquisition and autonomous problem-solving in a complex virtual world.
https://github.com/MineDojo/Voyager
This project provides a simulation framework for generative agents that mimic believable human behaviors within an interactive game environment, enabling running, saving, replaying, and demonstrating agent-based simulations.
SFighterAI is a deep reinforcement learning-based AI agent designed to master and beat the final boss in Street Fighter II: Special Champion Edition using only game screen pixel data.
A comprehensive Godot framework for creating smooth and complex AI motion using steering behaviors in both 2D and 3D games.
WorldX turns a single sentence into a running simulated world, generating maps and characters with LLMs and letting autonomous agents act, remember and form stories inside it.
Control room for LLM agent societies: pause and scrub a running simulation, question individual residents, inject instructions into the next step, and export replayable experiment packs.
PyGame Learning Environment (PLE) is a Python-based reinforcement learning platform that provides a variety of games and tools to facilitate the development and testing of reinforcement learning algorithms.
A locally runnable, low-cost implementation of generative agents simulating human-like behavior in an interactive Dungeons & Dragons setting using large language models.
GPTRPG is a proof-of-concept project featuring a GPT-based AI agent that autonomously interacts within a simple RPG-like environment using the OpenAI API.