Ann-Sophie Page

SCP workflow analysis 117 There are however some limitations. First, there may be scenarios where phases occur with an unknown number of repetitions, which would render our formulation unsuitable. Our TRN method is not suitable for real-time application, since it requires navigating the video in arbitrary temporal order. TRN may have scalability issues, since we need to train a different agent for each phase, which may be impractical if a very large number of phases is considered. This could potentially be alleviated by expanding the multi-agent framework to handle multiple phase transitions simultaneously. TRN is also sensitive to agent initialization, and while we propose 2 working strategies (FI, RMI), they can potentially be further optimized.

RkJQdWJsaXNoZXIy MTk4NDMw