“…Some of these (limiting) assumptions can be summarized as follows: (1) each respondent is represented by a single tuple in the microdata table; (2) all data to be released are stored in a single table; (3) once released, data are not further modified; (4) all the data that need to be released are available to the data holder before their release; (5) the same degree of privacy is guaranteed to all data respondents; (6) the released microdata table has a single quasi-identifier, known in advance; and (7) no external knowledge (except for that behind linking attacks counteracted by kanonymity) is available to recipients. Recently, the scientific community has started to extend the pioneering techniques illustrated so far in this chapter removing these assumptions, proposing solutions specifically tailored for supporting, among other scenarios: (1) multiple tuples per respondent (e.g., [101,107]); (2) release of multiple tables (e.g., [86,107]); (3) data republication (e.g., [113]); (4) continuous data release (e.g., [71,109,118]); (5) personalized privacy preferences (e.g., [56,112]); (6) multiple and/or non-predefined quasi-identifiers (e.g., [89,101]); (7) adversarial external knowledge (e.g., [21,76,79]). Figure 2.4 summarizes some notable solution recently proposed to extend the definitions of k-anonymity,`-diversity, and tcloseness removing the above-illustrated assumptions.…”