The increasing complexity of software systems has heightened the need for efficient and accurate vulnerability detection. Large Language Models have emerged as promising tools in this domain; however, their reasoning capabilities and limitations remain insufficiently explored. This study presents a systematic evaluation of different Large Language Models with and without explicit reasoning mechanisms, including Claude-3.5-Haiku, GPT-4o-Mini, DeepSeek-V3, O3-Mini, and DeepSeek-R1. Experimental results demonstrate that reasoning-enabled models, particularly DeepSeek-R1, outperform their non-reasoning counterparts by leveraging structured step-by-step inference strategies and valuable reasoning traces. With proposed data processing and prompt design in the interaction, DeepSeek-R1 achieves an accuracy of 0.9507 and an F1-score of 0.9659 on the Software Assurance Reference Dataset. These findings highlight the potential of integrating reasoning-enabled Large Language Models into vulnerability detection frameworks to simultaneously improve detection performance and interpretability.
Since the official launch of domestic exchange-traded options in 2015, options have gradually become an important component of the financial market. With the diversification of option types and the expansion of market size, option pricing research has received widespread attention. In this paper, we propose a finite difference method based on exponential integrals for the pricing of European call options, based on the Black-Scholes differential equation. Through numerical analysis, this method discretizes only the price region and uses the exponential Euler method to solve the nonhomogeneous system of linear differential equations in the time direction. Numerical experiments have verified the effectiveness of this method, showing that it can stably solve option pricing problems, especially when the price is close to the exercise price, demonstrating superior numerical performance.
Large language models (LLMs) have shown great potential in solving complex mathematical problems, but their performance in university-level mathematics is still underexplored. This study provides a systematic evaluation of eleven state-of-the-art LLMs on five core undergraduate mathematics courses. An end-to-end automated pipeline is proposed for solution inference and answer evaluation, including a reliable ensemble evaluation scheme using seven reasoning-enabled LLMs as expert evaluators. The experimental results show that reasoning models outperform non-reasoning ones significantly, with DeepSeek-V3.2 and Kimi-K2.5 achieving average scores of 88.24 and 88.99, respectively. Under the conditions of the conducted experiment in this study, DeepSeek-V3.2 achieves the most reasonable accuracy–cost trade-off. This work reveals the strengths and limitations of modern LLMs in advanced mathematical reasoning and provides insights into their application in AI mathematical education.
An odd $ c $-coloring of a graph is a proper $ c $-coloring such that each non-isolated vertex has at least one color appearing an odd number of times in its neighborhood. The minimum number of colors in any odd coloring of $ G $, denoted $ \chi_o(G) $, is called the odd chromatic number. This concept was introduced by Petruševski and Škrekovski, who conjectured that every planar graph $ G $ is odd 5-colorable and observed that $ \chi_o(G \square H) \leq \chi_o(G) \cdot \chi_o(H) $ for connected nontrivial graphs $ G $ and $ H $. In this paper, for specific Cartesian product graphs $ G \square H $, such as $ P_m \square P_n $, $ C_m \square P_n $, and $ C_m \square C_n $, we determine the exact value of $ \chi_o(G \square H) $, which establishes tighter upper bounds than the multiplicative bound $ \chi_o(G) \cdot \chi_o(H) $. We show that $ \chi_o(P_m \square P_n) \leq 4 $ with a complete characterization of all cases; $ \chi_o(C_m \square P_n) \leq 5 $ with a full classification for even and odd $ m $; and $ \chi_o(C_m \square C_n) \leq 5 $ with necessary and sufficient conditions for 3-, 4-, and 5-colorability under parity and divisibility constraints. These results significantly improve upon the multiplicative upper bound and provide new constructive methods and theoretical insights for studying odd colorings in Cartesian product graphs. Additionally, we determine that $ \chi_o(K_m\square P_n) = \chi_o(K_m \square C_n) = m $ for $ m = 3 $ and $ C_n $ is an even cycle or $ m \geq 4 $.