IP Library › Granted Patent US 12,242,959
Granted Patent B2
US 12,242,959 · App. 17/202,721 · Granted Mar 4, 2025

Method, apparatus, device and storage medium for embedding user app interest

Inventors: Huiqiang Zhong (Beijing, CN); Siqi Xu (Beijing, CN); Chenhui Liu (Beijing, CN); Lianghui Chen (Beijing, CN); Jun Fang (Beijing, CN)
Assignee: Beijing Baidu Netcom Science and Technology Co., Ltd.
G06N3/08G06F8/61G06F16/2282G06F16/2477G06N3/04
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 12,242,959
App. No.
17/202,721
Granted
Mar 4, 2025
Kind
B2
Abstract

A method, an apparatus, a device and a storage medium for embedding user app interest are provided. The method includes: acquiring a user existing app installation list and a user app installation list within a predetermined time window, where the app includes app ID information and app category information; inputting the existing app installation list and the app installation list within the predetermined time window into a pre-trained user app interest embedding model to obtain a user app interest embedding vector. By combining the user existing app installation list information and the user recent app installation list information, the user app interest embedding vector may simultaneously reflect the user long-term interest and the user short-term interest.

Claims (67)

1. A method for embedding user app interest, the method comprising:

acquiring a user existing app installation list and a user app installation list within a predetermined time window, wherein an app comprises app ID information and app category information; and

inputting the user existing app installation list and the user app installation list within the predetermined time window into a pre-trained user app interest embedding model to obtain a user app interest embedding vector, wherein the pre-trained user app interest embedding model is a neural network, and is trained by:

collecting a set of sample user app installation lists comprising a sample current user app installation list, and a sample previous user app installation list within a sample predetermined time window before a sample current time point, wherein the sample current user app installation list comprises first app ID information and first app category information, and the sample previous user app installation list comprises second app ID information and second app category information;

generating a first sample app embedding vector of the sample current user app installation list and generating a second sample app embedding vector of the sample previous user app installation list, by randomly setting an app ID query vector table and an app category query vector table and splicing the app ID query vector table and the app category query vector table; representing the sample current user app installation list as a first one-hot feature, representing the sample previous user app installation list as a second one-hot feature, determining a first input vector of the sample current app installation list by multiplying the first sample app embedding vector with the first one-hot feature, and determining a second input vector of the sample previous app installation list by multiplying the second sample app embedding vector with the second one-hot feature; and

training the pre-trained user app interest embedding model using the first input vector and the second input vector as an input.

2. The method according to claim 1 , wherein training the pre-trained user app interest embedding model using the first input vector and the second input vector comprises:

inputting the input vectors into the pre-trained user app interest embedding model;

constructing a decoding network model corresponding to the pre-trained user app interest embedding model, and using an output of the pre-trained user app interest embedding model as an input of the decoding network model; and

performing back propagation, based on a decoding error of the sample current user app installation list by the decoding network model and a decoding error of the sample previous user app installation list by the decoding network model, and adjusting the app ID query vector table and the app category query vector table, so that the pre-trained user app interest embedding model converges.

3. The method according to claim 1 , wherein,

the acquiring a user existing app installation list and a user app installation list within a predetermined time window, further comprises:

periodically acquiring the user existing app installation list and the user app installation list within the predetermined time window according to a periodic time window, wherein the app comprises the app ID information and the app category information;

the inputting the user existing app installation list and the user app installation list within the predetermined time window into the pre-trained user app interest embedding model to obtain a user app interest embedding vector, further comprises:

inputting the periodically acquired user existing app installation list and the user app installation list within the predetermined time window into the pre-trained user app interest embedding model to obtain the user app interest embedding vector corresponding to the periodic time window; and

the method further comprises:

storing the user app interest embedding vector corresponding to the periodic time window into a key-value library.

4. The method according to claim 1 , wherein collecting the set of sample user app installation lists comprising the sample current user app installation list and the sample previous user app installation list comprises:

collecting a user app installation log, and associating app installation information in the app installation log with time;

counting a number of users of the training app, and selecting a specific number of app sets by screening according to the number of users of the training app;

determining the sample current user app installation list and the sample previous user app installation list according to the app sets; and

acquiring the first app ID information, the first app category information, second app ID information and second app category information from the sample current user app installation list and the sample previous user app installation list.

5. The method according to claim 1 , wherein the pre-trained user app interest embedding model is an autoencoder network model.

6. An electronic device, comprising:

at least one processor; and

a memory communicating with the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions, when executed by the at least one processor, cause the at least one processor to perform operations comprising:

acquiring a user existing app installation list and a user app installation list within a predetermined time window, wherein an app comprises app ID information and app category information; and

inputting the user existing app installation list and the user app installation list within the predetermined time window into a pre-trained user app interest embedding model to obtain a user app interest embedding vector, wherein the pre-trained user app interest embedding model is a neural network, and is trained by:

collecting a set of sample user app installation lists comprising a sample current user app installation list, and a sample previous user app installation list within a sample predetermined time window before a sample current time point, wherein the sample current user app installation list comprises first app ID information and first app category information, and the sample previous user app installation list comprises second app ID information and second app category information;

generating a first sample app embedding vector of the sample current user app installation list and generating a second sample app embedding vector of the sample previous user app installation list, by randomly setting an app ID query vector table and an app category query vector table and splicing the app ID query vector table and the app category query vector table; representing the sample current user app installation list as a first one-hot feature, representing the sample previous user app installation list as a second one-hot feature, determining a first input vector of the sample current app installation list by multiplying the first sample app embedding vector with the first one-hot feature, and determining a second input vector of the sample previous app installation list by multiplying the second sample app embedding vector with the second one-hot feature; and

training the pre-trained user app interest embedding model using the first input vector and the second input vector as an input.

7. The electronic device according to claim 6 , wherein training the pre-trained user app interest embedding model using the first input vector and the second input vector comprises:

inputting the input vectors into the pre-trained user app interest embedding model;

constructing a decoding network model corresponding to the pre-trained user app interest embedding model, and using an output of the pre-trained user app interest embedding model as an input of the decoding network model; and

performing back propagation, based on a decoding error of the sample current user app installation list by the decoding network model and a decoding error of the sample previous user app installation list by the decoding network model, and adjusting the app ID query vector table and the app category query vector table, so that the pre-trained user app interest embedding model converges.

8. The electronic device according to claim 6 , wherein,

the acquiring a user existing app installation list and a user app installation list within a predetermined time window, further comprises:

periodically acquiring the user existing app installation list and the user app installation list within the predetermined time window according to a periodic time window, wherein the app comprises the app ID information and the app category information;

the inputting the user existing app installation list and the user app installation list within the predetermined time window into the pre-trained user app interest embedding model to obtain a user app interest embedding vector, further comprises:

inputting the periodically acquired user existing app installation list and the user app installation list within the predetermined time window into the pre-trained user app interest embedding model to obtain the user app interest embedding vector corresponding to the periodic time window; and

storing the user app interest embedding vector corresponding to the periodic time window into a key-value library.

9. The electronic device according to claim 6 , wherein collecting the set of sample user app installation lists comprising the sample current user app installation list and the sample previous user app installation list comprises:

collecting a user app installation log, and associating app installation information in the app installation log with time;

counting a number of users of the training app, and selecting a specific number of app sets by screening according to the number of users of the training app;

determining the sample current user app installation list and the sample previous user app installation list according to the app sets; and

acquiring the first app ID information, the first app category information, second app ID information and second app category information from the sample current user app installation list and the sample previous user app installation list.

10. A non-transitory computer readable storage medium storing computer instructions, wherein the computer instructions when executed by a computer, cause the computer to perform operations comprising:

acquiring a user existing app installation list and a user app installation list within a predetermined time window, wherein an app comprises app ID information and app category information; and

inputting the user existing app installation list and the user app installation list within the predetermined time window into a pre-trained user app interest embedding model to obtain a user app interest embedding vector, wherein the pre-trained user app interest embedding model is a neural network, and is trained by:

collecting a set of sample user app installation lists comprising a sample current user app installation list, and a sample previous user app installation list within a sample predetermined time window before a sample current time point, wherein the sample current user app installation list comprises first app ID information and first app category information, and the sample previous user app installation list comprises second app ID information and second app category information;

generating a first sample app embedding vector of the sample current user app installation list and generating a second sample app embedding vector of the sample previous user app installation list, by randomly setting an app ID query vector table and an app category query vector table and splicing the app ID query vector table and the app category query vector table; representing the sample current user app installation list as a first one-hot feature, representing the sample previous user app installation list as a second one-hot feature, determining a first input vector of the sample current app installation list by multiplying the first sample app embedding vector with the first one-hot feature, and determining a second input vector of the sample previous app installation list by multiplying the second sample app embedding vector with the second one-hot feature; and

training the pre-trained user app interest embedding model using the first input vector and the second input vector as an input.

11. The storage medium according to claim 10 , wherein training the pre-trained user app interest embedding model using the first input vector and the second input vector comprises:

inputting the input vectors into the pre-trained user app interest embedding model;

constructing a decoding network model corresponding to the pre-trained user app interest embedding model, and using an output of the pre-trained user app interest embedding model as an input of the decoding network model; and

performing back propagation, based on a decoding error of the sample current user app installation list by the decoding network model and a decoding error of the sample previous user app installation list by the decoding network model, and adjusting the app ID query vector table and the app category query vector table, so that the pre-trained user app interest embedding model converges.

12. The storage medium according to claim 10 , wherein,

the acquiring a user existing app installation list and a user app installation list within a predetermined time window, further comprises:

periodically acquiring the user existing app installation list and the user app installation list within the predetermined time window according to a periodic time window, wherein the app comprises the app ID information and the app category information;

the inputting the user existing app installation list and the user app installation list within the predetermined time window into the pre-trained user app interest embedding model to obtain a user app interest embedding vector, further comprises:

inputting the periodically acquired user existing app installation list and the user app installation list within the predetermined time window into the pre-trained user app interest embedding model to obtain the user app interest embedding vector corresponding to the periodic time window; and

storing the user app interest embedding vector corresponding to the periodic time window into a key-value library.

13. The storage medium according to claim 10 , wherein collecting the set of sample user app installation lists comprising the sample current user app installation list and the sample previous user app installation list comprises:

collecting a user app installation log, and associating app installation information in the app installation log with time;

counting a number of users of the training app, and selecting a specific number of app sets by screening according to the number of users of the training app;

determining the sample current user app installation list and the sample previous user app installation list according to the app sets; and

acquiring the first app ID information, the first app category information, second app ID information and second app category information from the sample current user app installation list and the sample previous user app installation list.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Mar 16, 2021
From: ZHONG, HUIQIANG; XU, SIQI; LIU, CHENHUI; CHEN, LIANGHUI; FANG, JUN
To: BEIJING BAIDU NETCOM SCIENCE AND TECHNOLOGY CO., LTD.
Reel/Frame 055605/0191 →
Priority Claims (1)
CN 202010995356.9 · Sep 21, 2020 · national
Continuity (1)
Related Publication 20210201149A1 · Jul 1, 2021
References Cited (41)
US 10873782B2 · Lomada · 2020 [cited by examiner]
US 11122333B2 · Li · 2021 [cited by examiner]
US 11681756B2 · Lee · 2023 [cited by examiner]
US 20100332326A1 · Ishai · 2010 [cited by applicant]
US 20130117329A1 · Bank et al. · 2013 [cited by applicant]
US 20190080225A1 · Agarwal et al. · 2019 [cited by applicant]
US 20200107072A1 · Lomada · 2020 [cited by examiner]
US 20200132491A1 · Zhang et al. · 2020 [cited by applicant]
US 20200374589A1 · Li · 2020 [cited by examiner]
US 20210034677A1 · Lee · 2021 [cited by examiner]
CA 2997797A1 · 2019 [cited by applicant]
CN 107729560A · 2018 [cited by applicant]
CN 108062342A · 2018 [cited by applicant]
CN 110765352A · 2020 [cited by applicant]
CN 111242748A · 2020 [cited by applicant]
CN 111523007A · 2020 [cited by applicant]
KR 20200007900A · 2020 [cited by applicant]
Remo Manuel Frey, Runhua Xu, Christian Ammendola, Omar Moling, Giuseppe Giglio, Alexander Ilic, Mobile recommendations based on interest prediction from consumer's installed apps-insights from a large-scale field study,… [cited by examiner]
Cheng, V.C., Chen, L., Cheung, W.K et al. A heterogeneous hidden Markov model for mobile app recommendation. Knowl Inf Syst 57, 207-228 (2018). (Year: 2018). [cited by examiner]
Huoran Li, Xuan Lu, Xuanzhe Liu, Tao Xie, Kaigui Bian, Felix Xiaozhu Lin, Qiaozhu Mei, and Feng Feng. 2015. Characterizing Smartphone Usage Patterns from Millions of Android Users. In Proceedings of the 2015 Internet Me… [cited by examiner]
Zhao, Sha, Gang Pan, Yifan Zhao, Jianrong Tao, Jinlai Chen, Shijian Li, and Zhaohui Wu. “Mining user attributes using large-scale app lists of smartphones.” IEEE Systems Journal 11, No. 1: 315-323. (Year: 2016). [cited by examiner]
Li, Tong, Tong Xia, Huandong Wang, Zhen Tu, Sasu Tarkoma, Zhu Han, and Pan Hui. “Smartphone app usage analysis: datasets, methods, and applications.” IEEE Communications Surveys & Tutorials 24, No. 2 (2022): 937-966. (Y… [cited by examiner]
Li, D., & Song, D.H. (2017). Detecting Low Rating Android Apps Before They Have Reached the Market. ArXiv, abs/1712.05843. ( Year: 2017). [cited by examiner]
M. Rungta, P. P. Sherki, M. P. Dhaliwal, H. Tiwari and V. Vala, “Two-Phase Multimodal Neural Network for App Categorization using APK Resources,” 2020 IEEE 14th International Conference on Semantic Computing (ICSC), San… [cited by examiner]
K. Katsarou, G. Yu and F. Beierle, “WhatsNextApp: LSTM-Based Next-App Prediction With App Usage Sequences,” in IEEE Access, vol. 10, pp. 18233-18247, 2022, (Year: 2022). [cited by examiner]
Q. Ma, S. Muthukrishnan and W. Simpson, “App2Vec: Vector modeling of mobile apps and applications,” 2016 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM), San Francisco, CA, … [cited by examiner]
Q. Ma, “Modeling Users for Online Advertising,” Thesis, Rutgers, State Univ. New Jersey, Oct. 2016, 115 total pages. (Year: 2016). [cited by examiner]
Lee, Jinwoo, Yena Lee, Ming Jin, John Kim, and Jiman Hong. “Analysis of application installation logs on android systems.” In Proceedings of the 34th ACM/SIGAPP Symposium on Applied Computing, pp. 2140-2145. 2019. (Year… [cited by examiner]
Zhao, S., Xu, F., Xu, Y et al. Investigating smartphone user differences in their application usage behaviors: an empirical study. CCF Trans. Pervasive Comp. Interact. 1, 140-161 (2019). (Year: 2019). [cited by examiner]
S. Zhao et al., “AppUsage2Vec: Modeling Smartphone App Usage for Prediction,” 2019 IEEE 35th International Conference on Data Engineering (ICDE), Macao, China, 2019, pp. 1322-1333 (Year: 2019). [cited by examiner]
Zhang, Junqi, Bing Bai, Ye Lin, Jian Liang, Kun Bai and Fei Wang. “General-Purpose User Embeddings based on Mobile App Usage.” Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mi… [cited by examiner]
E. H.-C. Lu, Y.-W. Lin and J.-B. Ciou, “Mining mobile application sequential patterns for usage prediction,” 2014 IEEE International Conference on Granular Computing (GrC), Noboribetsu, Japan, 2014, pp. 185-190 (Year: 2… [cited by examiner]
Zhung-Xun Liao, Yi-Chin Pan, Wen-Chih Peng, and Po-Ruey Lei. 2013. On mining mobile apps usage behavior for predicting apps usage in smartphones. In Proceedings of the 22nd ACM international conference on Information & … [cited by examiner]
C. Chen, T. Maekawa, A. Daichi and T. Hara, “Preliminary Investigation of Predicting Next-Use Mobile Apps Using App Semantic Representations,” 2019 IEEE International Conference on Pervasive Computing and Communications… [cited by examiner]
Hua-Dong Yue, “Determining The Recommended Application APP Method and Device”, published on Dec. 2, 2015, Document ID: CN-105117440A, pp. 12 (Year: 2015). [cited by examiner]
An-teng Pan, “Using The Recommendation Method, Device, Storage Medium and Computer Equipment”, published on Nov. 6, 2018, Document ID: CN 108769125 A, pp. 15 (Year: 2018). [cited by examiner]
Jia et al., “Method, Device, Device and Storage Medium for Training Tag Generation Model”, published on Aug. 20, 2024, but filed on Aug. 31, 2020, Document ID: CN 112036486 B, pp. 25 (Year: 2024). [cited by examiner]
Xu et al., “A context-based using a sequence of APP recommending method and system”, published on Mar. 23, 2018, Document ID: CN-107832426-A, pp. 13 (Year: 2018). [cited by examiner]
Zhou et al., “APP starting accelerating method based on heterogeneous network with attribute embedded”, published on Jul. 7, 2020, Document Id: CN-111381902-A, p. 45 (Year: 2020). [cited by examiner]
European Patent Office, Extended European Search Report for Application No. 21162110.7, dated Sep. 7, 2021 (26 pages). [cited by applicant]
Baltrunas, L et al.; “Frappe: Understanding the Usage and Perception of Mobile App Recommendations In-the-Wild”; ARXIV.org, Cornell University Library, 201 Olin Library, Cornell University, Ithaca, New York 14853; May 1… [cited by applicant]