Tokyo-based startup Colorful Board released a fashion coordination app called Sensy on 7 November. The app will “learn” (store in memory) an individual’s fashion preference using artificial intelligence and propose outfits to one’s likings.
The Sensy app shows several items, and one can choose “like” or “not like” for each of these. By continuing this process, the artificial intelligence in the app learns one’s preference and generally improves a recommendation’s accuracy. If one finds any favorite fashion item to buy in the recommendation, a person may easily jump into an online store via the app selling that item. The company has allied with over 1,600 fashion brands including those which have not yet been introduced in Japan, so users can gain recommendations regarding a wide variety of available items.
Through the machine learning process, the app will come to obtain a fashion sense similar to the one an individual has. By adding “sophistication” from fashion models and stylists to it, the app will assist in the selection of appropriate items, called the Sense Link function.
Colorful Board has been developing the core technology using artificial intelligence in partnership with Keio University Professor Dr. Eitaro Aiyoshi and Chiba University Professor Dr. Takashi Okamoto. Patents for this technology are pending in the U.S.
Conventional recommendation systems like those adopted on Amazon uses the collaborative filtering approach to produce recommendations, which makes proposals based on preference sets of those who are similar to a user in profile. In contrast, the Sensy app will make recommendations based on one’s preference only so the method will work properly even when recommending brand new or niche items, which are unlikely to be found in the preference sets of other users.