Mathematical reasoning ability objectively reflects a language model's understanding of implicit knowledge in contexts, with logic being a prerequisite for exploring, articulating and establishing effective reasoning. Large language models (LLMs) have shown great potential in complex reasoning tasks represented by mathematical reasoning. However, existing mathematical datasets either focus on commonsense reasoning, assessing the model's knowledge application ability, or arithmetic problems with fixed calculation rules, evaluating the model's rapid learning capability. There is a lack of datasets that require solving problems solely through logical reasoning. As a result, the performance of LLMs in accurately understanding the implicit logical relationships in problems and deriving conclusions based solely on given conditions is hindered. To address this challenge, we construct a dataset specifically for multiple step reasoning tasks: Reasoning-Math (RMath). This dataset focuses on evaluating logical reasoning abilities with mathematical reasoning problems, covering typical problem types, including direct reasoning problems, hypothetical reasoning problems, and nested reasoning problems. Additionally, we design a standardized annotation scheme that transforms natural language descriptions of conditions into formal propositions. Other annotation contents include problem categories, proposition truth values, and proposition relationship types. This not only reduces biases caused by semantic misunderstandings during problem-solving, but also facilitates the incorporation of theoretically grounded logical reasoning methods to enhance reasoning abilities. Furthermore, we propose a normalization problem-solving framework based on propositional logic for RMath and design the problem-solving process for prompt tuning to guide LLMs to absorb mathematical logical theories and improving reasoning abilities. Finally, we evaluate RMath on several popular LLMs and present the corresponding results.

As a convenient means of transportation, unmanned aerial vehicles (UAVs) can provide consumers and merchants with safe, efficient, and contactless instant delivery services. However, online orders are often concurrent, dynamic, and with time constraints in an instant delivery system. Therefore, it's important to optimally schedule the delivery sequence upon the UAVsystem. In this paper, we design a real-time UAV delivery scheduling model considering dynamic online orders. Through our instant delivery management system, the urban delivery network can be updated in real time according to order information. Considering the practical situation, customers with spatial diversity may produce orders with common order requirements, so we consider the overlap of UAV routes in the scheduling process. An order merging algorithm (OMA) is then proposed to improve the delivery efficiency (the ratio of the number of completed orders to the times UAVs visit stores) of UAVs. With the help of proposed system-level decision-making method, our system can meet real-time concurrent order demands across urban delivery networks. To evaluate the performance of system, we further carry out simulation experiments to verify the effectiveness of our scheme. Results show that the proposed UAVscheduling scheme improves the efficiency of UAV instant delivery.

When flying robots are used in close-range interaction with humans, the noise they generate, also called consequential sound, is a critical parameter for user acceptance. We conjecture that there is a benefit in adding natural sounds to noisy domestic drones. To test our hypothesis experimentally, we carried out a mixed-methods research study (N = 56) on reported user perception of a sonified domestic flying robot with three sound conditions at three distances. The natural sounds studied were, respectively, added to the robot’s inherent noises during flying; namely, a bird song and a rain sound, plus a control condition of no added sound. The distances studied were set according to proxemics; namely, near , middle , and far . Our results show that adding bird song or rain sound affects the participants’ perceptions, and the proxemic distances play a nonnegligible role. For instance, we found that participants liked the bird condition the most when the drone was at far , while they disliked the same sound the most when at near . We also found that participants’ perceptions strongly depended on their associations and interpretations deriving from previous experience. We derived six concrete design recommendations.