High performance and low complexity channel decoding algorithms have always been one of the core technologies in 6G mobile communication. Polar codes are the first type of constructive coding with achievable capacity, and are the forefront direction of channel coding research. For the traditional SC (successful cancellation) decoding algorithm, when the Polar code length is short, there is a problem of incomplete channel polarization that affects the accuracy of decoding, which cannot satisfy the requirements of ultra-high data rate, ultra-low latency, and ultra-low power consumption for communication in the 6G era. Deep learning methods have become a research hotspot due to their outstanding performance in decoding performance. Through the data-driven approach, neural networks were applied to the decoding of Polar codes, and the impact of different dataset construction methods on the decoding performance of the network was compared. Using the concept of sequence classification, a (16,8) short Polar code decoder was proposed innovatively based on the cascaded residual neural network (Residual Network) and gate recurrent unit (GRU). Under the same signal-to-noise ratio, the Polar code decoder based on Res-GRU has lower bit error rate and frame error rate compared to traditional SC decoding algorithm, and has lower decoding delay.

This study investigates how interactive dashboards influence decision making by exploring how specific dashboard features impact design task performance, efficiency, understanding, and confidence. An experiment was conducted in which undergraduate student participants were given a design activity and randomly assigned to one of five dashboards, each using the same underlying functions but varying in the visualization features employed. These features include different graphical representations of the design decision inputs and performance outputs. Participants were first asked to use their assigned dashboard to design a catapult system that maximizes launch distance while meeting requirements related to height, weight, and cost. Following the design task, they were asked a series of questions about their experiences with the dashboard and their understanding of the catapult model. A between-subjects analysis then evaluated how the dashboard design influenced various outcomes of interest. The results show that students who used the most feature-rich dashboard did not perform objectively better than those with the most feature-sparse dashboard, though their self-reported performance was higher. The performance of female versus male participants was also compared, with no significant differences found. The findings support the notion that dashboards should be designed with minimal features to convey the necessary information, and they also point out the disconnect between objective performance and user-assessed performance with interactive dashboards.