Semantic Textual Similarity (STS) measures the degree of semantic equivalence between two texts.
Paraphrase identification is the task of recognizing text fragments with approximately the same meaning within a specific context.
For more detail, see my quick and dirty webpages: