Tools and libraries

1. AI GYM, 2023–2026. The goal of this library is to provide a set of tools for learning the basics of Artificial Intelligence. In other words, it helps you learn how to build agents that solve problems through search. It implements several search algorithms and a common interface for agents, allowing you to create and deploy problem-solving agents easily.


2. CARRETE, L. F.; CASTANARES, M.; DAMIANI, E.; BARTH, F.. DSSE: Drone Swarm Search Environment. 2023. The Drone Swarm Search project is a PettingZoo-based environment for use with multi-agent or single-agent reinforcement learning algorithms. Agents (drones) must find targets (shipwreck survivors) without knowing their positions or receiving rewards based on their distance to them. Instead, agents receive the probabilities of targets being in each map cell. The project supports the study of reinforcement learning algorithms that require dynamic probabilities as inputs.


3. ANDRADE, J.; SILVA, W.; SANTOS, G.; BARTH, F.. MLBugDetection. 2022. Machine learning explainability and unexpected-behavior detection. Most explainability packages require both trained models and training data to create explainer objects. This package supports ceteris paribus feature analysis using only a trained model and one or more input samples.