Fabian M. Suchanek Rule Mining 140 >skip intro
Semantic IE You are here 2 Source Selection and Preparation Entity Recognition Entity Disambiguation singer Fact Extraction Reasoning Instance Extraction singer Elvis
Overview •  Introduction •  Rules •  Inductive Logic Programming •  Confidence in a KB •  Top-Down Rule Mining •  Bottom-Up Rule Mining   3
Reminder: IE by reasoning (H@1, “married”,R@ 1)[1] (“married',  )[1] (H@1,  )[5] (R@1,  )[2] (R@1,  )[3] [10] [6] + Pattern Application Rule [10] loses 2 wins 3 loses 6 “Hermione married Ron” “Ron” spouse World1: spouse World2: spouse loses 3 wins 6 wins 2
Reminder: IE by reasoning “Hermione married Ron” spouse Where do we get these rules from? (H@1, “married”,R@ 1)[1] (“married',  )[1] (H@1,  )[5] (R@1,  )[2] (R@1,  )[3] [10] [6] + Pattern Application Rule [10]
© TV week © Dailymail © Public Domain