Ambiguity

• Unlike formal languages (like C), natural language is ambiguous.
• Words have multiple meanings (e.g. run, can) and multiple syntactic classes (note, can, run).
• There is ambiguity in sentences (E.g. I saw the girl with the telescope).
• There is referential ambiguity (what does it refer to?)
• Peole don't usually see this ambiguity, but it is really hard for computers.
• When parsing, we usually use semantics to resolve the ambiguity.
• We have all kinds of prior knowledge that help us resolve the ambiguities.
• These are big issues in NLP.
• While the technology is not perfect, it can be used today.