ASMR: Agentic Schema Generation for Ship Maintenance Report Writing

2026-07-09

AI Summary

This work addresses the challenge of automatically discovering concise yet informative template structures from multiclass historical ship maintenance and operational reports. To this end, the authors propose the ASMR framework, which introduces, for the first time, a collaborative multi-agent mechanism to this task. The framework comprises two core components: a field-generation agent that extracts semantic concepts from text via adaptive multi-granularity clustering and produces candidate fields, and a structure-optimization agent that employs reinforcement learning to select a compact, non-redundant, and highly informative final template. Experimental results demonstrate that the generated templates effectively guide the composition of more complete, consistent, and actionable maintenance reports, thereby validating the method’s effectiveness and novelty in automated template discovery.

Abstract

In this paper, we study the automatic schema generation problem: given a collection of historical ship maintenance and operational reports across multiple form categories, automatically discover compact and informative schemas that capture the essential information requirements of each report type. To address this challenge, we propose ASMR, a modular agentic framework consisting of two specialized agents. A Field Generation Agent extracts semantic concepts from historical narratives and generates candidate schema fields through adaptive multi-granularity clustering, while a Structural Optimizer Agent employs reinforcement learning to identify compact, informative, and non-redundant schema representations. The resulting schemas can guide report authors toward producing more complete, consistent, and actionable reports. Preliminary results demonstrate the promise of the proposed approach and highlight several open research challenges at the intersection of data management, agentic AI, and human-centered AI.

Problem

• schema generation
• ship maintenance reports
• automatic schema discovery
• information extraction
• report standardization

Innovation

• agentic AI
• schema generation
• multi-granularity clustering
• reinforcement learning
• human-centered AI