IP Library Granted Patent US 12,633,232
Granted Patent B2
US 12,633,232 · App. 17/845,315 · Granted May 19, 2026

Deep learning-based pedagogical word recommendation method for predicting and improving vocabulary skills of foreign language learners

Inventors: June Young Park (Yongin-si, KR); Jae Min Shin (Seoul, KR)
Assignee: Socra AI Inc.
G09B19/06G09B5/02G06F3/04845
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,633,232
App. No.
17/845,315
Granted
May 19, 2026
Kind
B2
Abstract

According to an aspect of the present specification, a method in which a terminal recommends a word to a user includes: receiving recommended word information from a server, wherein the recommended word information includes word information that the user is predicted not to know in an AI (Artificial Intelligence) model of the user on the basis of training data of the user; displaying a first window including the word information on the basis of the recommended word information; receiving operation of dragging the first window from the user; displaying an icon representing whether to add the word information to a vocabulary list of the terminal on the basis of a direction of the dragging operation; and including the word information in the vocabulary on the basis of the direction of the dragging operation.

Claims (44)

1 . A word recommendation method performed by a terminal, the method comprising:

receiving recommended word information from a server, wherein the recommended word information includes word information that a user is predicted not to know in an AI (Artificial Intelligence) model of the user based on training data of the user;

displaying a first window including the word information based on the recommended word information;

receiving operation of dragging the first window from the user;

displaying an icon representing whether to add the word information to a vocabulary list of the terminal based on a direction of the dragging operation; and

including the word information in the vocabulary list based on the direction of the dragging operation,

wherein the AI model includes (1) a user embedding model for generating an embedding vector of the user, (2) a word embedding model for generating an embedding vector of a word related to the word information, and (3) an arbitrary similarity scoring function of determining similarity between the user embedding vector acquired from the user embedding model and the word embedding vector acquired from the word embedding model,

wherein the server inputs (1) the user embedding vector and (2) the word embedding vector to the arbitrary similarity scoring function and outputs a result value for predicting whether the user knows a word related to the word vector from the arbitrary similarity scoring function,

wherein the recommended word information is generated in the server based on the result value, and

wherein the method further comprises:

displaying a word, a first button and a second button on the terminal,

based on the first button being selected by the user, marking the word displayed on the terminal as “0” indicating that the user knows the word displayed on the terminal,

based on the second button being selected by the user, marking the word displayed on the terminal as “1” indicating that the user does not know the word displayed on the terminal,

based on the word displayed on the terminal being marked as “0”, displaying meaning of the word displayed on the terminal, a third button and a fourth button on the terminal,

based on the third button being selected by the user, maintaining a marking “0” indicating that the user knows the word displayed on the terminal, and

based on the fourth button being selected by the user, changing the marking “0” to “1” indicating that the user does not know the word displayed on the terminal.

2 . The word recommendation method according to claim 1 , further comprising updating the training data based on the vocabulary list and transmitting the training data to the server.

3 . The word recommendation method according to claim 1 , wherein the word information is included in the vocabulary list when the direction of the dragging operation is a first direction, and the word information is not included in the vocabulary list when the direction of the dragging operation is a direction opposite to the first direction.

4 . The word recommendation method according to claim 3 , wherein in the displaying the icon, the word information is moved in the direction of the dragging operation in response to the dragging operation, and movement of the icon is displayed in accordance with the word information.

5 . The word recommendation method according to claim 1 , further comprising:

displaying a second window including the word information of the vocabulary list based on the vocabulary list; and

receiving touch operation of the second window for checking meaning of the word related to the word information based on the word information from the user.

6 . The word recommendation method according to claim 5 , further comprising displaying a third window for presenting meaning of the word information of the vocabulary list based on a point of receiving the touch operation.

7 . The word recommendation method according to claim 6 , wherein the displaying the third window is to rotate and display the second window in a direction corresponding to the point of receiving the touch operation.

8 . A terminal which performs a word recommendation method, the terminal comprising:

a communication module;

a memory;

a display unit; and

a processor,

wherein the processor receives recommended word information from a server through the communication module, the recommended word information includes word information that a user is predicted not to know in an AI (Artificial Intelligence) model based on training data of the user, and the processor allows the display unit to display a first window including the word information based on the recommended word information, to receive operation of dragging the first window from the user, and to display an icon representing whether to add the word information to a vocabulary list included in the memory based on a direction of the dragging operation, and includes the word information in the vocabulary list based on the direction of the dragging operation,

wherein the AI model includes (1) a user embedding model for generating an embedding vector of the user, (2) a word embedding model for generating an embedding vector of a word related to the word information, and (3) an arbitrary similarity scoring function of determining similarity between the user embedding vector acquired from the user embedding model and the word embedding vector acquired from the word embedding model,

wherein the server inputs (1) the user embedding vector and (2) the word embedding vector to the arbitrary similarity scoring function, outputs a result value for predicting whether the user knows a word related to the word vector from the arbitrary similarity scoring function,

wherein the recommended word information is generated in the server based on the result value, and

wherein the processor:

displays a word, a first button and a second button on the terminal,

based on the first button being selected by the user, marking the word displayed on the terminal as “0” indicating that the user knows the word displayed on the terminal,

based on the second button being selected by the user, marks the word displayed on the terminal as “1” indicating that the user does not know the word displayed on the terminal,

based on the word displayed on the terminal being marked as “0”, displays meaning of the word displayed on the terminal, a third button and a fourth button on the terminal,

based on the third button being selected by the user, maintains a marking “0” indicating that the user knows the word displayed on the terminal, and

based on the fourth button being selected by the user, changes the marking “0” to “1” indicating that the user does not know the word displayed on the terminal.

9 . The terminal according to claim 8 , wherein the processor updates the training data based on the vocabulary list, and transmits the training data to the server through the communication module.

10 . The terminal according to claim 8 , wherein the processor allows the display unit to display a second window including the word information of the vocabulary list based on the vocabulary list, and to receive touch operation of the second window for checking meaning of the word related to the word information based on the word information from the user.

11 . The terminal according to claim 10 , wherein the processor allows the display unit to display a third window for presenting meaning of the word information of the vocabulary list based on a point of receiving the touch operation.

12 . The word recommendation method according to claim 1 , wherein the result value is 0 when the user related to the user embedding vector is predicted to know the word related to the word embedding vector, and the result value is 1 when the user related to the user embedding vector is predicted not to know the word related to the word embedding vector.

Assignments (3)
CHANGE OF NAME Recorded Feb 27, 2026
From: RIIID INC.
To: SOCRA AI INC.
Reel/Frame 075025/0276 →
CHANGE OF NAME Recorded Feb 18, 2026
From: RIIID INC.
To: SOCRA AI INC.
Reel/Frame 074906/0011 →
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Jun 21, 2022
From: PARK, JUNE YOUNG; SHIN, JAE MIN
To: RIIID INC.
Reel/Frame 060263/0888 →
Priority Claims (2)
KR 10-2021-0080065 · Jun 21, 2021 · national
KR 10-2022-0031245 · Mar 14, 2022 · national
Continuity (1)
Related Publication 20220406216A1 · Dec 22, 2022
References Cited (11)
US 10564792B2 · Kim · 2020 [cited by examiner]
US 20220121826A1 · Ma · 2022 [cited by examiner]
US 20220406217A1 · Park · 2022 [cited by examiner]
CN 107346183B · 2021 [cited by examiner]
CN 109271636B · 2023 [cited by examiner]
KR 1020170080519A · 2017 [cited by applicant]
KR 1020190082453A · 2019 [cited by applicant]
KR 102213476B1 · 2021 [cited by applicant]
KR 1020210057954A · 2021 [cited by applicant]
CN107346183B Translation (Year: 2017). [cited by examiner]
CN109271636B Translation (Year: 2018). [cited by examiner]