Tool-using large language models (LLMs) have improved at producing executable API calls, but reliable tool use requires coordinating semantic intent with executable invocation, especially in multi-step settings. Existing tool-learning objectives and traces collapse these two roles into a single homogeneous generation stream, so the learning signal cannot distinguish a flawed plan from a flawed call. We propose Dual-Mode Unified Policy Optimization (DUPO), a typed trajectory optimization framework that represents each step with a PLAN span for the current subgoal and a TOOL span for the API invocation. DUPO follows a PLAN-to-TOOL rollout factorization, optimizes the two modes with mode-separated policy ratios, and introduces Chain-Aware Upstream Advantage, so that early planning errors, which propagate through the rest of the trajectory, receive stronger credit than later steps. To support this optimization, we introduce PlanArena, a human-annotated dataset of 20,646 trajectories across 7,941 scenarios with explicit PLAN–TOOL supervision. Experiments on BFCL, API-Bank, and Tau2 show that DUPO improves tool-use performance, with gains concentrated in multi-step settings. Our code and dataset will be released upon publication.

While large language models (LLMs) perform strongly after adaptation within individual scientific domains, their learning dynamics in multi-disciplinary contexts remain poorly understood, despite the promise of improved generalization and broader applicability through cross-domain knowledge synergy. We present a systematic study of multi-discipline LLM fine-tuning across five scientific disciplines and compare the learning dynamics of full fine-tuning, LoRA, LoRA-MoE, and LoRA composition. In our naturally imbalanced setting, we show that multi-disciplinary learning is substantially more variable than single-discipline training and distill four empirical patterns: (1) Balance-then-Diversity: increasing scarce-domain exposure helps, but broad source-pool diversity must be preserved; (2) Merge-then-Align: general instruction data restores instruction-following behavior needed for effective joint adaptation; (3) Optimize-then-Scale: increasing trainable capacity offers limited gains without prior design optimization; and (4) Share-then-Specialize: asymmetric LoRA-MoE improves multi-discipline performance through a shared low-rank projection. Together, these patterns form a practical recipe for principled multi-discipline fine-tuning and provide actionable guidance for developing generalizable scientific LLMs.

Multi-agent systems have emerged as a promising approach for enhancing the reasoning capabilities of large language models in complex problem-solving. However, current MAS frameworks are limited by poor flexibility and scalability, with underdeveloped optimization strategies. To address these challenges, we propose ReSo, which integrates task graph generation with a reward-driven two-stage agent selection process. The core of ReSo is the proposed Collaborative Reward Model, which can provide fine-grained reward signals for MAS cooperation for optimization. We also introduce an automated data synthesis framework for generating MAS benchmarks, without human annotations. Experimentally, ReSo matches or outperforms existing methods. ReSo achieves 33.7% and 32.3% accuracy on Math-MAS and SciBench-MAS SciBench, while other methods completely fail. The code and data are available at Reso.

Extractive text summarisation aims to select salient sentences from a document to form a short yet informative summary. While learning-based methods have achieved promising results, they have several limitations, such as dependence on expensive training and lack of interpretability. Therefore, in this paper, we propose a novel non-learning-based method by for the first time formulating text summarisation as an Optimal Transport (OT) problem, namely Optimal Transport Extractive Summariser (OTExtSum). Optimal sentence extraction is conceptualised as obtaining an optimal summary that minimises the transportation cost to a given document regarding their semantic distributions. Such a cost is defined by the Wasserstein distance and used to measure the summary’s semantic coverage of the original document. Comprehensive experiments on four challenging and widely used datasets - MultiNews, PubMed, BillSum, and CNN/DM demonstrate that our proposed method outperforms the state-of-the-art non-learning-based methods and several recent learning-based methods in terms of the ROUGE metric.