ACT: Agentic Classification Tree

Method and contribution

Learns classification trees of natural-language binary questions through impurity-based split selection and TextGrad refinement.

A later agentic formulation of learning question-based decision trees; compare question optimization and tree construction.

Read in context

LLM decision trees: from Tree Prompting to agentic classification →

Citation

Vincent Grari, Tim Arni, Thibault Laugel, Sylvain Lamprier, James Zou, Marcin Detyniecki. ACT: Agentic Classification Tree. arXiv:2509.26433.

@misc{act,
  title = {{ACT: Agentic Classification Tree}},
  author = {Grari, Vincent and Arni, Tim and Laugel, Thibault and Lamprier, Sylvain and Zou, James and Detyniecki, Marcin},
  year = {2025},
  url = {https://arxiv.org/html/2509.26433v4},
  eprint = {2509.26433},
  archivePrefix = {arXiv}
}