Reviews: Multilingual Anchoring: Interactive Topic Modeling and Alignment Across Languages

The paper proposes an anchor-based approach for learning a topic model in multilingual settings. The model first is built automatically and then can be refined by interaction with a user. The paper addresses a problem of interest of the previous NIPS submissions. Quality: The paper is technically sound. I like that for bilingual topic modeling it is only require to have a dictionary between two languages but documents should not be aligned for training a topic model.