DARA: Decomposition-Alignment-Reasoning Autonomous Language Agent for Question Answering over Knowledge Graphs

Answering Questions over Knowledge Graphs (KGQA) is key to well-functioning autonomous language agents in various real-life applications.To improve the neural-symbolic reasoning capabilities of language agents powered by Large Language Models (LLMs) in KGQA, we propose the Decomposition-Alignment-Reasoning Agent (DARA) framework.DARA effectively parses questions into formal queries through a dual mechanism: highlevel iterative task decomposition and low-level task grounding.Importantly, DARA can be efficiently trained with a small number of highquality reasoning trajectories.Our experimental results demonstrate that DARA fine-tuned on LLMs (e.g.Llama-2-7B, Mistral) outperforms both in-context learning-based agents with GPT-4 and alternative fine-tuned agents, across different benchmarks in zero-shot evaluation.This makes such models more accessible for real-life applications.We also show that DARA attains performance comparable to state-of-the-art enumerating-and-rankingbased methods for KGQA 1 .1 Code, model are released at https://github.com/ UKPLab/acl2024-