![[2602.12574] Monte Carlo Tree Search with Reasoning Path Refinement for Small Language Models in Conversational Text-to-NoSQL](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.12574] Monte Carlo Tree Search with Reasoning Path Refinement for Small Language Models in Conversational Text-to-NoSQL
Summary
This paper presents a novel framework, Stage-MCTS, which enhances small language models' ability to generate NoSQL queries through conversational interactions, outperforming existing models in accuracy.
Why It Matters
As NoSQL databases become integral to various sectors, simplifying their querying process is crucial. This research addresses the challenge of conversational query generation, making it more accessible for users without technical expertise, thus broadening the usability of NoSQL databases.
Key Takeaways
- Introduces the Conversational Text-to-NoSQL task for generating queries from dialogue.
- Presents Stage-MCTS, which combines Monte Carlo Tree Search with reasoning path refinement.
- Constructs CoNoSQL, a dataset of 2,000+ dialogues for evaluation.
- Demonstrates improved execution value match accuracy by up to 7.93% over existing models.
- Addresses the gap in current research focusing on multi-turn conversational queries.
Computer Science > Databases arXiv:2602.12574 (cs) [Submitted on 13 Feb 2026] Title:Monte Carlo Tree Search with Reasoning Path Refinement for Small Language Models in Conversational Text-to-NoSQL Authors:Xubang Xiong, Raymond Chi-Wing Wong, Yuanfeng Song View a PDF of the paper titled Monte Carlo Tree Search with Reasoning Path Refinement for Small Language Models in Conversational Text-to-NoSQL, by Xubang Xiong and 2 other authors View PDF HTML (experimental) Abstract:NoSQL databases have been widely adopted in big data analytics, geospatial applications, and healthcare services, due to their flexibility and scalability. However, querying NoSQL databases requires specialized technical expertise, creating a high barrier for users. While recent studies have explored text-to-NoSQL problem, they primarily focus on single-turn interactions, ignoring the conversational nature of real-world queries. To bridge this gap, we introduce the Conversational Text-to-NoSQL task, which generates NoSQL queries given a natural language question, a NoSQL database, and the dialogue history. To address this task, we propose Stage-MCTS, a framework that endows small language models (SLMs) with NoSQL-specific reasoning capabilities by formulating query generation as a search problem. The framework employs Monte Carlo Tree Search (MCTS) guided by a rule-based reward to produce stepwise reasoning data, followed by progressive supervised fine-tuning (SFT) and self-training strategies. We further c...
