2 results about "Multi perspective" patented technology

The invention provides a question and answer method based on multi-agent collaborative reasoning, and the method comprises the steps: simulating a multi-angle debate and consensus forming process of a human expert team through configuring a plurality of agents of different professional analysis roles; taking a complete dialogue inference summary, a structured response plan and a response text as a triple training data set, selecting a basic language model as a student model, and performing instruction fine tuning and optimization on the student model by using the triple training data set; and inputting the question text of the user into the optimization model to obtain a final response text. The invention further discloses a method for generating high-quality (question, plan and answer) data pairs by using the method, and performing fine adjustment on the miniaturized model according to the high-quality (question, plan and answer) data pairs. According to the method, the complex multi-agent reasoning and planning capability can be'distilled 'into a single model with lower deployment cost, the problem that the complex model is difficult to deploy on a large scale in practical application is solved, and the industrial practicability and economical efficiency of the technology are improved.

This invention provides a traffic flow prediction method and system based on multi-perspective spatiotemporal modeling, belonging to the field of traffic information management technology. This invention decouples the spatiotemporal modeling process into two complementary perspectives: implicit spatiotemporal modeling and globally explicit spatiotemporal relationships. Based on a selective state-space model architecture, it performs in-depth analysis of input features from both channel and spatial location perspectives, thereby constructing a multi-perspective temporal understanding path. Furthermore, since the implicit and globally explicit temporal modeling capture corresponding temporal information from the channel and spatial location perspectives respectively, a channel space attention module is introduced to effectively filter out unimportant temporal information interference, further improving the overall model's traffic flow prediction capability. The method described in this invention not only effectively improves the accuracy of traffic flow prediction, but its design concept can also provide valuable insights for other spatiotemporal data-driven downstream tasks (such as weather forecasting).