Search papers, labs, and topics across Lattice.
This paper introduces StructPO, a struct-aware policy learning framework that streamlines the generation of academic paper introductions by integrating the multi-stage writing process into a single-pass policy using explicit stage tokens. By employing struct-aware credit assignment and refinement-guided optimization, StructPO effectively enhances semantic alignment, structural rationality, and inference efficiency compared to traditional workflow-based approaches. The framework not only generalizes well to out-of-domain settings but also competes favorably with GPT-5.1 in human evaluations when scaled to Qwen3-32B, highlighting its potential as a cost-effective alternative for academic writing tasks.
StructPO achieves a coherent academic introduction in a single pass, outperforming traditional multi-stage workflows while maintaining competitive quality against advanced LLMs.
Generating a rigorous paper introduction with large language models (LLMs) remains challenging, since it requires coordinating background, gap identification, method and contribution within a coherent narrative. Existing solutions externalize this process as multi-stage prompts or agent workflows which are expensive and vulnerable to cross-stage drift. We propose StructPO, a struct-aware policy learning framework that internalizes the entire multi-stage writing workflow into a single-pass policy controlled by explicit stage tokens. StructPO introduces struct-aware credit assignment to decouple local stage quality from global coherence and refinement-guided optimization to internalize revision behavior into the first-pass policy. Experiments show that StructPO improves semantic alignment, structural rationality and inference efficiency over workflow-based baselines, generalizes to out-of-domain settings, and remains competitive with GPT-5.1 in human evaluation when scaled to Qwen3-32B. These results show that internalizing academic writing workflows through fine-grained policy optimization offers a viable alternative to costly external orchestration.
