Yakun Zhu, Yutong Huang, Shengqian Qin, Zhongzhen Huang, Shaoting Zhang, Xiaofan Zhang. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Large language models (LLMs) have showcased remarkable capabilities as autonomous agents when augmented with external tools.Equipped with fixed tool sets, LLMs struggle with addressing diverse user inquiries in open-world tasks.To evaluate and boost the performance of LLMs in dealing with complex demands in the real-world, we propose open-world function calling, where LLMs need to retrieve suitable tools from a pre-defined external tool library and use retrieved tools to resolve the user's problem.We introduce Meta-Tool, a versatile and plug-and-play tool retrieval system as the access of LLMs to external tool library.Drawing inspiration from the myriad of enhanced approaches associated with Retrieval-Augmented Generation (RAG), Meta-Tool employs a hypothesizeretrieve-invoke framework.We further propose Meta-Bench, a comprehensive benchmark for evaluating LLMs in open-world function calling and associated tasks.Meta-Bench encompasses 2, 800 dialogues and 7, 361 tools, spanning ten distinct scenarios to provide robust and diverse test categories.In conjunction, we present MT-LLaMA, a finetuned version of LLaMA-3.1, which exhibits remarkable performance improvements.Our empirical experiments reveal that Meta-Tool significantly enhances the ability of advanced LLMs to retrieve and leverage the most suitable tools compared to previous tool retrieval methods.Moreover, our fine-tuning enables even smallersized LLMs to achieve comparable even exceeding results to GPT-4o.Both the benchmark and the model are made publicly available at https://github.com/qinshengqian/Meta-Tool to foster further research and development in the field.

Existing context-based offline meta-reinforcement learning (COMRL) methods primarily focus on task representation learning and given-context adaptation performance. They often assume that the adaptation context is collected using task-specific behavior policies or through multiple rounds of collection. However, in real applications, the context should be collected by a policy in a one-shot manner to ensure efficiency and safety. We find that intrinsic context ambiguity across multiple tasks and out-of-distribution (OOD) issues due to distribution shift significantly affect the performance of one-shot adaptation, which has been largely overlooked in most COMRL research. To address this problem, we propose using heteroscedastic uncertainty in representation learning to identify ambiguous and OOD contexts, and train an uncertainty-aware context collecting policy for effective one-shot online adaptation. The proposed method can be integrated into various COMRL frameworks, including classifier-based, reconstrution-based and contrastive learning-based approaches. Empirical evaluations on benchmark tasks show that our method can improve one-shot adaptation performance by up to 36% and zero-shot adaptation performance by up to 34% compared to existing baseline COMRL methods.

Yakun Zhu, Shaohang Wei, Xu Wang, Kui Xue, Shaoting Zhang, Xiaofan Zhang. Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2025.

In most of the recent literature about model predictive control (MPC) on vehicles, the finite-time and bipartite control are rarely considered. Although existing MPC methods can be applied to address the separation and rendezvous problem, they may suffer from time wasting and consistency conflicts. The finite-time control facilitates the motion planning by pre-calculating the coverage time horizon, which can avoid the unnecessary time-wasting. The bipartite control can realize the collision avoidance for the multi-vehicle system unless the reference state is zero. Therefore, a finite-time bipartite (FTB) control method for the UAV-USV system based on MPC is proposed. However, in most of the existing UAV-USV systems, the sustainable operating mechanism is not considered as usual. In this paper, to realize the sustainable operating mechanism, an UAV-USV ecosystem is introduced and further improved. Specifically, a battery replacement process is proposed and realized by the MPC method with switching topologies. Considering an UAV-USV system, the UAVs on the USV can help the USV replace the rechargeable batteries. And this process can be realized automatically by the proposed MPC method with switching topologies in a finite time. The recursive feasibility and asymptotic stability of the proposed method are proven. In addition, a series of simulations are given to demonstrate the effectiveness and advantages of the proposed method.