This document discusses personalized influential topic search via social network summarization. It proposes extracting representative users from a social network as a social summarization for a given topic, to measure topics' influence at a similar scale. Two random walk based approaches are developed for effective topic-aware social summarization: random clustering and an L-length random walk. This social summarization can then be used to efficiently find the top-k most influential topics related to a user's query. The approaches are evaluated on real-world datasets and shown to be effective and efficient for personalized influential topic search in social networks.