AI-Based Intelligent Decision-Making System Integrating Multi-Modal Large Model

By integrating text, image, audio and other modal data, this paper makes use of the powerful data processing and feature extraction capabilities of the large model to improve the accuracy and comprehensiveness of the decision-making system. During research and verification, the multi-modal data fusion strategy, the selection and training of large models, and the construction and evaluation of the intelligent decision system (IDS) are discussed. The system architecture design includes six levels: data acquisition, preprocessing, feature extraction, modal fusion, large model processing and decision output. Modular design is adopted to support the input and processing of various modal data, and the fusion strategy and model parameters can be flexibly adjusted according to specific tasks. In the multi-modal data fusion strategy, this paper proposes a mid-term fusion approach, which dynamically adjusts the importance weight of each modal feature through a multi-modal attention network to optimize the fusion effect. In the experiment, the multi-modal data sets, including emotion analysis, event detection, and object recognition, are used to evaluate the system's ability in data fusion, large model accuracy, and intelligent decision-making accuracy. The results show that IDS with a multi-modal large model performs well in emotion analysis, event detection, and object recognition, with accuracy reaching 87.5%, 84.7%, and 90.4%, respectively, significantly higher than the traditional methods.