@article{reinforcementlearningforcompositionalgen,
  title = {Reinforcement Learning for Compositional Generalization with Outcome-Level Optimization},
  author = {Xiyan Fu and Wei Liu},
  year = {2026},
  eprint = {2605.04920},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2605.04920},
}