arXiv:2509.24230v2 Announce Type: replace Abstract: Large Language Models (LLMs) enable intelligent multi-robot collaboration but face fundamental trade-offs: open-loop methods that compile tasks into formal representations for external executors produce sound plans but lack adaptability in partially observable environments, while iterative methods incur prohibitive computational costs that scale poorly with team size and task complexity. In this paper, we propose Efficient Long-Horizon Planning (ELHPlan), a novel framework that introduces Action Chains, sequences of actions explicitly bound to sub-goal intentions, as the fundamental planning primitive. ELHPlan operates via a cyclical process: 1) constructing intention-bound action sequences, 2) proactively validating for conflicts and feasibility, 3) refining issues through targeted mechanisms, and 4) executing validated actions. This design balances adaptability and efficiency by providing intention-bound action sequences with longer lookahead while avoiding expensive full re-planning. We further advocate comprehensive efficiency metrics, including token consumption and planning time, to more holistically evaluate multi-agent collaboration. Our experiments on benchmarks TDW-MAT and C-WAH demonstrate that ELHPlan achieves comparable task success rates while consuming only 30-40% of the tokens required by state-of-the-art methods. Our research establishes a new efficiency-effectiveness frontier for LLM-based multi-agent planning systems.

arXiv:2506.18485v3 Announce Type: replace-cross Abstract: Reinforcement Learning with Verifiable Rewards~(RLVR) has emerged as a powerful learn-to-reason paradigm for large reasoning models to tackle complex tasks. However, the current RLVR paradigm is still not efficient enough, as it works in a trial-and-error manner. To perform better, the model needs to explore the reward space by numerously generating responses and learn from fragmented reward signals, blind to the overall reward patterns. Fortunately, verifiable rewards make the natural language description of the reward function possible, and meanwhile, LLMs have demonstrated strong in-context learning ability. This motivates us to explore if large reasoning models can benefit from a \textbf{motivation} of the task, \textit{i.e.}, awareness of the reward function, during the reinforcement finetuning process, as we humans sometimes do when learning. In this paper, we introduce \textit{\textbf{M}otivation-\textbf{e}nhanced \textbf{R}einforcement \textbf{F}inetuning}~(\textbf{MeRF}), an intuitive yet effective method enhancing reinforcement finetuning of LLMs by involving \emph{``telling LLMs rules of the game''}. Specifically, \textbf{MeRF} directly injects the reward specification into the prompt, which serves as an in-context motivation for the model to be aware of the optimization objective. This simple modification leverages the in-context learning ability of LLMs, aligning generation with optimization, thereby incentivizing the model to generate desired outputs from both inner motivation and external reward. Empirical evaluations demonstrate that \textbf{MeRF} achieves substantial performance gains over the RLVR baseline. Moreover, ablation studies show that MeRF performs better with greater consistency between the in-context motivation and the external reward function, while the model also demonstrates an ability to adapt to misleading motivations through reinforcement finetuning.

arXiv:2603.03099v2 Announce Type: replace-cross Abstract: Despite Adam demonstrating faster empirical convergence than SGD in many applications, much of the existing theory yields guarantees essentially comparable to those of SGD, leaving the empirical performance gap insufficiently explained. In this paper, we uncover a key second-moment normalization in Adam and develop a stopping-time/martingale analysis that provably distinguishes Adam from SGD under the classical bounded variance model (a second moment assumption). In particular, we establish the first theoretical separation between the high-probability convergence behaviors of the two methods: Adam achieves a $\delta^{-1/2}$ dependence on the confidence parameter $\delta$, whereas corresponding high-probability guarantee for SGD necessarily incurs at least a $\delta^{-1}$ dependence.