Abstract
Techniques are described for automated uniform resource locator (URL) routing template inference to enable faster web automation, robotic process automation (RPA), and direct deep link generation. Tool indexer pipeline 110 receives web address dataset 100 representing collected links from a specific website. Path decomposition module 112 parses web address dataset 100 into ordered segments, and tabular mapping module 114 constructs a table organizing these segments. Entropy calculation module 116 and hierarchical clustering module 118 recursively partition the table to algorithmically filter out random variables and isolate consistent structural website patterns. Pattern extraction module 120 identifies static website paths and parametric wildcard positions to produce structural routing template 128. Language model module 132 receives web page context buffer 130 and appends semantic descriptors (meaningful labels) to structural routing template 128 to generate automated tool definition 134 for browser automation agent 160 navigation, ultimately improving the speed and reliability of AI-assisted internet browsing.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Awano, Yuki; Ray, Aniket; Gora, Surya; and Arbab Yazd, Sara, "FREQUENT ROUTE INFERENCE BY ENTROPY SLICING FOR AUTOMATED DIRECT LINK SYNTHESIS", Technical Disclosure Commons, ()
https://www.tdcommons.org/dpubs_series/12293