The Google Operating Systems blog has slides from Peter Norvig's presentation at UC Berkeley on how the Google One Box Q&A results work. He says that "Google doesn't use predefined patterns, they find the patterns from examples, as this approach is more scalable." The slides show the algorithms that detect these patterns from examples. Is it perfect? No. Ben at the Search Engine Roundtable discovered Google OneBox Q&A Adult Spam last month, in which Matt Cutts confirmed to be an issue with the OneBox extractor code.
Scroll to top