The mario-ai project offers a comprehensive pipeline for developing AI agents to master Super Mario Bros. using neuroevolution. By integrating a Python-based NEAT implementation with the OpenAI Gym SuperMario environment, it allows users to define custom fitness criteria, mutation rates, and network topologies. During training, the framework evaluates generations of neural networks, selects high-performing genomes, and provides real-time visualization of both gameplay and network evolution. Additionally, it supports saving and loading trained models, exporting champion genomes, and generating detailed performance logs. Researchers, educators, and hobbyists can extend the codebase to other game environments, experiment with evolutionary strategies, and benchmark AI learning progress across different levels.

MARTI (Multi-Agent Reinforcement learning Toolkit and Interface) is a research-oriented framework that streamlines the development, evaluation, and benchmarking of multi-agent RL algorithms. It offers a plug-and-play architecture where users can configure custom environments, agent policies, reward structures, and communication protocols. MARTI integrates with popular deep learning libraries, supports GPU acceleration and distributed training, and generates detailed logs and visualizations for performance analysis. The toolkit’s modular design allows rapid prototyping of novel approaches and systematic comparison against standard baselines, making it ideal for academic research and pilot projects in autonomous systems, robotics, game AI, and cooperative multi-agent scenarios.

Reinforcement Learning Agents for PettingZoo Games is a Python-based code library delivering off-the-shelf DQN, PPO, and A2C algorithms for multi-agent reinforcement learning on PettingZoo environments. It features standardized training and evaluation scripts, configurable hyperparameters, integrated TensorBoard logging, and support for both competitive and cooperative games. Researchers and developers can clone the repo, adjust environment and algorithm parameters, run training sessions, and visualize metrics to benchmark and iterate quickly on their multi-agent RL experiments.

BomberManAI is an AI agent designed to play the classic Bomberman game autonomously. Developed in Python, it interfaces with a game environment to perceive map states, available moves, and opponent positions in real time. The core algorithm combines A* pathfinding, breadth-first search for reachability analysis, and a heuristic evaluation function to determine optimal bomb placement and evasion strategies. The agent handles dynamic obstacles, power-ups, and multiple opponents on various map layouts. Its modular architecture enables developers to experiment with custom heuristics, reinforcement learning modules, or alternative decision-making strategies. Ideal for game AI researchers, students, and competitive bot developers, BomberManAI provides a flexible framework for testing and improving autonomous gaming agents.

Java Action Generic is a lightweight, modular library that allows developers to implement autonomous agent behaviors in Java by defining generic actions. Actions are parameterized units of work that agents can execute, schedule, and compose at runtime. The framework offers a consistent action interface, allowing developers to create custom actions, handle action parameters, and integrate with LightJason’s agent lifecycle management. With support for event-driven execution and concurrency, agents can perform tasks such as dynamic decision-making, interaction with external services, and complex behavior orchestration. The library promotes reusability and modular design, making it suitable for research, simulations, IoT, and game AI applications on any JVM-supported platform.

PommerLearn enables researchers and developers to train multi-agent RL bots in the Pommerman game environment. It includes ready-to-use implementations of popular algorithms (PPO, DQN), flexible configuration files for hyperparameters, automatic logging and visualization of training metrics, model checkpointing, and evaluation scripts. Its modular architecture makes it easy to extend with new algorithms, customize environments, and integrate with standard ML libraries such as PyTorch.

VMAS is a comprehensive toolkit for building and training multi-agent systems using deep reinforcement learning. It supports GPU-based parallel simulation of hundreds of environment instances, enabling high-throughput data collection and scalable training. VMAS includes implementations of popular MARL algorithms like PPO, MADDPG, QMIX, and COMA, along with modular policy and environment interfaces for rapid prototyping. The framework facilitates centralized training with decentralized execution (CTDE), offers customizable reward shaping, observation spaces, and callback hooks for logging and visualization. With its modular design, VMAS seamlessly integrates with PyTorch models and external environments, making it ideal for research in cooperative, competitive, and mixed-motive tasks across robotics, traffic control, resource allocation, and game AI scenarios.

DQN-Deep-Q-Network-Atari-Breakout-TensorFlow provides a complete implementation of the DQN algorithm tailored for the Atari Breakout environment. It uses a convolutional neural network to approximate Q-values, applies experience replay to break correlations between sequential observations, and employs a periodically updated target network to stabilize training. The agent follows an epsilon-greedy policy for exploration and can be trained from scratch on raw pixel input. The repository includes configuration files, training scripts to monitor reward growth over episodes, evaluation scripts to test trained models, and TensorBoard utilities for visualizing training metrics. Users can adjust hyperparameters such as learning rate, replay buffer size, and batch size to experiment with different setups.

GPT or NPC integrates the powerful capabilities of generative AI to create dynamic non-player characters (NPCs) in games. This innovation allows NPCs to engage players in realistic conversations, adapt to various scenarios, and respond intelligently to player actions. By utilizing machine learning and natural language processing, the technology enhances the depth of storytelling and interactivity, making each gaming experience unique. Whether you're exploring medieval towns or battling creatures, GPT or NPC allows for engaging dialogues and personalized interactions, elevating the overall gaming experience.

Human or Not is an engaging AI-powered game that challenges players to discern whether their conversation partner is a human or an AI. Based on chatroulette, this game offers a fun way to test your ability to distinguish between human and machine interactions. Using advanced language models such as GPT-4 and AI21 Labs' Jurassic-2, it provides an intriguing and entertaining experience for all ages.