	US 12,277,591 B2
	Reducing sample selection bias in a machine learning-based recommender method, system, and non-transitory computer-readable medium
Yang Shi, San Mateo, CA (US); and Youngjoo Chung, San Mateo, CA (US)
Assigned to Rakuten Group, Inc., Tokyo (JP)
Filed by Rakuten Group, Inc., Tokyo (JP)
Filed on Jan. 28, 2022, as Appl. No. 17/588,099.
Claims priority of provisional application 63/221,872, filed on Jul. 14, 2021.
Prior Publication US 2023/0036964 A1, Feb. 2, 2023
Int. Cl. G06Q 30/0601 (2023.01); G06N 3/08 (2023.01)

CPC G06Q 30/0631 (2013.01) [G06N 3/08 (2013.01)]

9 Claims

1. A machine-learning method, performed by a computer system, for predicting user-item interaction values on an ecommerce platform that includes products from different shops with different sales volumes, the method comprising:

performing the following with respect to a training phase:

obtaining a machine-learning model for predicting user-item interactions, wherein the model was previously trained to predict user-item interactions based on user and item data on an ecommerce platform that includes different shops with different sales volumes;

modifying the model to reduce sample selection bias in favor of shops with larger sales volumes by performing the following:

(a) identifying a sample batch of shops on the ecommerce platform;

(b) obtaining a shop-specific training dataset for each shop, wherein each shop-specific training dataset has item data for items in the shop and user data including user-item interaction data for items in the shop;

applying the model to user and item data in the shop-specific training dataset to obtain predicted user and item interactions for the shop;

calculating a shop-specific loss between the predicted user and item interactions and the actual user and item interactions in the shop-specific training dataset; and

(d) calculating a global loss based on the shop-specific losses;

(e) calculating a global parameter adjustment for the model to reduce the global loss;

(f) creating an updated model by adjusting the parameters of the model using the global parameter adjustment; and

(g) repeating steps (c)-(f) for a number of iterations, wherein the updated model in a previous iteration becomes the model in the next iteration;

performing the following with respect to a prediction phase:

using the updated model to obtain user-item interaction value predictions on the ecommerce platform with respect to user and item pairs for which no interaction value is known, wherein using the updated model to obtain user-item interaction value predictions comprises:

applying a user neural network encoder to input user data to generate a user vector representation,

applying an item neural network encoder to input item data to obtain an item vector representation, wherein each of the user and item neural network encoders has a plurality of hidden layers,

applying a dot product module to calculate a dot product of the user and item vector representations to generate a user-item interaction score, and

applying a classification module using the user-item interaction score to predict the user-item interaction value.