The field of automata learning is moving towards more efficient and effective methods for learning complex systems. Researchers are developing new techniques for active learning, including algorithms for learning symbolic NetKAT automata and weighted automata over number rings. Compositional approaches are also being explored, with a focus on learning synchronous systems and refining global alphabets into component alphabets. These developments have the potential to significantly improve the scalability and accuracy of automata learning. Notable papers in this area include:

• Active Learning of Symbolic NetKAT Automata, which presents algorithms for learning different types of NetKAT automata and evaluates their applicability to realistic network configurations.
• Learning Weighted Automata over Number Rings, Concretely and Categorically, which develops a generic reduction procedure for active learning problems and presents an exact learning algorithm for weighted automata over number rings.
• Compositional Active Learning of Synchronous Systems through Automated Alphabet Refinement, which develops a general technique for compositional learning of synchronizing parallel systems with unknown decomposition.