[Purpose/Significance] From the perspective of multimodal data-driven, this article clarifies the connotation of multimodal data and security situational awareness, and constructs a multimodal data-driven security situational awareness framework, aiming to timely and accurately perceive, understand, predict, and make decisions on network data security threats, and implement precise policies based on scientific decision-making. [Method/Process] Based on the security intelligence theory, the WSR system methodology, and literature review, this article constructed a multimodal data-driven security situational awareness framework according to the logical hierarchy of “physical-rational-humanistic”. [Result/Conclusion] The framework starts from the physical level of information technology and hardware facilities, and emphasizes on the multimodal data-driven situational awareness process and mechanism of situation crisis detection, situation element extraction, situation understanding, and situation prediction at the rational level. Ultimately, it focuses on the sound multimodal data situational awareness collaborative governance mechanism and intelligent decision-making empowerment science application at the humanistic level.

[Purpose/Significance] This paper aims to explore the construction method of public security event graph based on multimodal data, excavate the evolutionary law and internal logic of multimodal public security events, achieve a deep analysis and understanding of public safety events and enhance the ability of event prevention and control. [Method/Process] This paper obtained multimodal public safety event data including official bulletins, news reports and journal articles. By constructing public security ontology, event extraction and event relationship extraction, the multimodal public security event graph was constructed, and the temporal evolution and causal logic analysis of public security events were carried out based on the event graph. [Result/Conclusion] The research shows that the public security event graph based on multimodal data can reveal the event logic of public security events and provide scientific basis for the prevention and response of public security events.

[Purpose/Significance] This study develops an AI agent model for multimodal data processing across varied tasks. By exploring the inherent value of multimodal data, the model enhances the support for decision-making processes. [Method/Process] This study designed a hybrid AI agent model that integrated both horizontal and vertical structures. Using the financial annual reports from thirty Dow Jones Industrial Average companies as a case study, this model facilitated intelligent investment research and analysis for the year 2023. The research compared the performance of different organizational structures of financial agents in executing three subtasks: analysis of company fundamentals, market sentiment, and risk and opportunity assessment. Decisions were made based on these analyses, leveraging large language models capable of processing multimodal data. [Result/Conclusion] The vertical agent structure excels in tasks requiring efficient, directional strategic analysis with clear hierarchical decision-making. Conversely, the horizontal structure is more effective in scenarios demanding collaboration, flexibility, and broad expertise. The hybrid model outperforms single-structure agents in terms of accuracy in the final investment decision, demonstrating its efficacy in optimizing task-specific agent configurations.