When exploring anime, it is not easy for users to express their preferred narrative or character-related features as explicit search conditions. This study presents ChASER (Character + Anime + Serendipitous + Exploratory + Retrieval), which is a character-centered exploratory search demo.
Using ChASER, users start from a familiar or favorite character and progressively navigate through similar characters across multiple anime titles. The system uses a character feature space constructed from character descriptions, with preprocessing designed to reduce potential dependence on title-specific expressions. At each exploration step, users examine similar characters, spoiler-free character descriptions, anime information, and feature terms related to character similarity, and then select a character of interest for the next query. The exploration path is retained as a map, and after exploration, associated feature terms are summarized to provide cues for reflecting on the exploration process and the user’s own preferences. This demo enables a character-as-query interaction that lets users expand their exploration from familiar characters to unfamiliar ones and anime titles without requiring detailed search conditions.
