| CPC G06F 16/90328 (2019.01) [G06F 16/24578 (2019.01); G06Q 30/0631 (2013.01)] | 13 Claims |
|
1. An information processing apparatus comprising:
at least one memory configured to store program code;
at least one processor configured to operate as instructed by the program code, the program code including:
prefix acquisition code configured to cause at least one of the at least one processor to acquire a prefix input on an EC (E-Commerce) site by a user;
context acquisition code configured to cause at least one of the at least one processor to acquire a context representing a feature relating to a search on the EC site by the user;
candidate acquisition code configured to cause at least one of the at least one processor to generate a plurality of search query candidates, based on the prefix;
reranking code configured to cause at least one of the at least one processor to rerank the plurality of search query candidates, using a result generated by inputting the plurality of search query candidates and the context to a natural language processing model trained using information relating to purchases on the EC site by the user;
training data generation code configured to cause at least one of the at least one processor to generate training data including a positive pair comprising information of two items among a plurality of purchased items purchased by the user, and a negative pair comprising information of two items with at least one being information of a non-purchased item not purchased by the user, among a plurality of items being sold on the EC site; and
training code configured to cause at least one of the at least one processor to:
generate feature vectors of the positive pair and feature vectors of the negative pair by inputting the information to the natural language processing model, and
train the natural language processing model to minimize a distance between the feature vectors of the positive pair and maximize a distance between the feature vectors of the negative pair in a common vector space.
|