ACT: Agentic Classification Tree
This addresses the need for trustworthy AI in regulated domains by offering a method that combines interpretability with performance on unstructured data, though it is incremental as it builds on existing decision-tree and LLM techniques.
The paper tackles the problem of making AI decisions transparent and interpretable for high-stakes settings by extending decision trees to handle unstructured text inputs, resulting in ACT matching or surpassing prompting-based baselines on text benchmarks while providing clear decision paths.
When used in high-stakes settings, AI systems are expected to produce decisions that are transparent, interpretable, and auditable, a requirement increasingly expected by regulations. Decision trees such as CART provide clear and verifiable rules, but they are restricted to structured tabular data and cannot operate directly on unstructured inputs such as text. In practice, large language models (LLMs) are widely used for such data, yet prompting strategies such as chain-of-thought or prompt optimization still rely on free-form reasoning, limiting their ability to ensure trustworthy behaviors. We present the Agentic Classification Tree (ACT), which extends decision-tree methodology to unstructured inputs by formulating each split as a natural-language question, refined through impurity-based evaluation and LLM feedback via TextGrad. Experiments on text benchmarks show that ACT matches or surpasses prompting-based baselines while producing transparent and interpretable decision paths.