US Patent 10491622 Automatic webpage change detection

An improved web crawler, associated method of crawling the Internet, and automatically detection of crawled webpage changes are provided. The method comprises obtaining a first version and a second version of the webpage; generating a first simhash of the first version of the webpage and a second simhash of the second version of the webpage; calculating, using a similarity hashing function having small output perturbations for small input perturbations, a probability that there are no differences between the first version of the webpage and the second version of the webpage; providing, to one or more researcher computers, the first version of the webpage and the second version of the webpage; based on input identifying a change in the webpage, updating a count of changes associated with the webpage; providing information about the change in the webpage in the second version of the webpage relative to the first version of the webpage as feedback to the crawler.