- AutoSlog: AutoSlog is an Information Extraction Algorithm developed in the early 90s that automatically built dictionaries based on slot-fillint rules (extraction patterns).
- Context:
- AutoSlog was trained on an annotated corpus
- Used simple linguistic rules.
- See: [[Extraction-Rule_Learning?]]; [[Riloff,_1993?]].
Notes
- Used a corpus of answer keys and texts from MUC-4 to compare filled templates with the sentences they were taken from.
- Compared a sentence's linguistic parse to a set of heuristic linguistic patterns
- Built a syntactic pattern to extract noun phrases similar to the slotfillers from the template.
- Given the slot-filler "bombed", for example, Auto-Slog might find a sentence with the phrase "public buildings were bombed" and associate the pattern <subject> passive-verb with it, extracting "public buildings" as a dictionary term.
References