Evaluating the graph comprehension and reasoning abilities of Large Language Models (LLMs) is challenging and often incomplete. Existing benchmarks focus primarily on pure graph understanding, lacking a comprehensive evaluation across all graph types and detailed capability definitions. This paper presents GraCoRe, a benchmark for systematically assessing LLMs’ graph comprehension and reasoning. GraCoRe uses a three-tier hierarchical taxonomy to categorize and test models on pure graph and heterogeneous graphs, subdividing capabilities into 10 distinct areas tested through 19 tasks. Our benchmark includes 11 datasets with 5,140 graphs of varying complexity. We evaluate four closed-source and eight open-source LLMs, conducting thorough analyses from both ability and task perspectives. Key findings reveal that OpenAI o1 model has amazing comprehension and reasoning capabilities, semantic enrichment enhances reasoning performance, node ordering impacts task success, and the ability to process longer texts does not necessarily improve graph comprehension or reasoning.

Chain-of-Thought (CoT) reasoning improves performance on complex tasks but introduces significant inference latency due to its verbosity.In this work, we propose Multiround Adaptive Chain-of-Thought Compression (MACC), a framework that leverages the token elasticity phenomenon-where overly small token budgets may paradoxically increase output length-to progressively compress CoTs via multiround refinement.This adaptive strategy allows MACC to dynamically determine the optimal compression depth for each input.Our method achieves an average accuracy improvement of 5.6% over state-of-the-art baselines, while also reducing CoT length by an average of 47 tokens and significantly lowering latency.Furthermore, we show that test-time performance-accuracy and token length-can be reliably predicted using interpretable features like perplexity and compression rate on training set.Evaluated across different models, our method enables efficient model selection and forecasting without repeated fine-tuning, demonstrating that CoT compression is both effective and predictable.Our code will be released in https://github.com/Leon221220/ MACC.

Graph-theoretic problems arise in real-world applications like logistics, communication networks, and traffic optimization.These problems are often complex, noisy, and irregular, posing challenges for traditional algorithms.Large language models offer potential solutions but face several challenges, including limited accuracy, input length constraints, and suboptimal algorithm selection.To address these challenges, we propose MA-GTS (Multi-Agent Graph Theory Solver), a multi-agent framework that decomposes these complex problems through agent collaboration.MA-GTS maps the implicitly expressed textbased graph data into clear, structured graph representations and dynamically selects the most suitable algorithm based on problem constraints and graph structure scale.We validate MA-GTS using the G-REAL dataset, a real-world-inspired graph theory dataset we created.Experimental results show that MA-GTS outperforms state-of-the-art methods in cost-effectiveness, accuracy, and scalability, achieving strong results on multiple benchmarks (G-REAL 93.6%, GraCoRe 96.9% NL-Graph 98.4%) with robust performance on both closed-and open-source base models.

Multi-Agent Systems (MAS) excel at accomplishing complex objectives through the collaborative efforts of individual agents. Among the methodologies employed in MAS, Multi-Agent Reinforcement Learning (MARL) stands out as one of the most efficacious algorithms. However, when confronted with the complex objective of Formation Control with Collision Avoidance (FCCA): designing an effective reward function that facilitates swift convergence of the policy network to an optimal solution. In this paper, we introduce a novel framework that aims to overcome this challenge. By giving large language models (LLMs) on the prioritization of tasks and the observable information available to each agent, our framework generates reward functions that can be dynamically adjusted online based on evaluation outcomes by employing more advanced evaluation metrics rather than the rewards themselves. This mechanism enables the MAS to simultaneously achieve formation control and obstacle avoidance in dynamic environments with enhanced efficiency, requiring fewer iterations to reach superior performance levels. Our empirical studies, conducted in both simulation and real-world settings, validate the practicality and effectiveness of our proposed approach.Project Website: https://macsclab.github.io/LLM_FCCA

Multiagent formation obstacle avoidance is a crucial research topic in the field of multiagent cooperative control, and deep reinforcement learning has shown remarkable potential in this domain. However, most existing studies are not fully distributed and often involve relatively simple scenarios. In this article, we propose an advanced method based on multiagent deep reinforcement learning to address formation and obstacle avoidance in dynamic obstacles environments. For handling complex environments with an unknown number of obstacles, we use long short-term memory (LSTM) networks to encode dynamic obstacles, thereby improving the efficiency of obstacle avoidance. Our method achieves formation and obstacle avoidance in scenarios with both dynamic and static obstacles, where agents coordinate through fully independent and autonomous decision-making. We utilize the multiagent proximal policy optimization (MAPPO) algorithm for centralized training and distributed execution, enhancing the agents' formation and obstacle avoidance capabilities in complex settings. Through simulation and real-world experiments, and by comparing with benchmark methods, we demonstrate significant improvements in formation effectiveness and obstacle avoidance success rates, showcasing the superiority and practicality of our proposed approach.

Quadrotor UAVs, renowned for their agility and versatility, are extensively utilized in a range areas. However, their inherent underactuated dynamic characteristics render them particularly vulnerable to external disturbances and systemic failures. To address this issue, our study introduces a hybrid control method tailored to combat the most prevalent types of drone failures—actuator faults. This innovative approach leverages reinforcement learning to enhance fault tolerance. Specifically, we employ reinforcement learning techniques to output compensatory control signals that bolster the core functionalities of the base controller. This integration aims to preserve the stability and continuity of mission-critical tasks even in the face of operational faults, thereby ensuring robust safety controls. We utilized the proximal policy optimization algorithm for the strategic training of our control systems. We test in both simulated environments and real-world scenarios was conducted to evaluate the efficacy of our control strategy under conditions of actuator failure. The results affirm that our method significantly enhances the safety and stability of drone operations, maintaining control integrity during rotor failures.

Power estimation is the basis of many hardware optimization strategies. However, it is still challenging to offer accurate power estimation at an early stage such as high-level synthesis (HLS). In this paper, we propose PowerGear, a graph-learning-assisted power estimation approach for FPGA HLS, which features high accuracy, efficiency and transferability. PowerGear comprises two main components: a graph construction flow and a customized graph neural network (GNN) model. Specifically, in the graph construction flow, we introduce buffer insertion, datapath merging, graph trimming and feature annotation techniques to transform HLS designs into graph-structured data, which encode both intra-operation micro-architectures and inter-operation interconnects annotated with switching activities. Furthermore, we propose a novel power-aware heterogeneous edge-centric GNN model which effectively learns heterogeneous edge semantics and structural properties of the constructed graphs via edge-centric neighborhood aggregation, and fits the formulation of dynamic power. Compared with on-board measurement, PowerGear estimates total and dynamic power for new HLS designs with errors of 3.60% and 8.81%, respectively, which outperforms the prior arts in research and the commercial product Vivado. In addition, PowerGear demonstrates a speedup of 4× over Vivado power estimator. Finally, we present a case study in which PowerGear is exploited to facilitate design space exploration for FPGA HLS, leading to a performance gain of up to 11.2%, compared with methods using state-of-the-art predictive models.