Multi Task Learning: A Survey and Future Directions

Taeho Lee; Junhee Seok

doi:10.1109/ICAIIC57133.2023.10067098

Multi Task Learning: A Survey and Future Directions

Taeho Lee, Junhee Seok

School of Electrical Engineering

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

3 Citations (Scopus)

Abstract

Multi-task learning (MTL) is a problem that must be applied in modern recommendation systems and is just as difficult. In the recent e-commerce advertising market, it is necessary to be able to predict not only the probability of users clicking, but also the probability of conversion and purchase. By predicting multi-task, it is possible to increase the accuracy of each task and optimize advertisements for various goals of advertisers. Traditional conversion rate (CVR) prediction models have difficulty learning because the number of conversions is too small compared to the total number of impressions. This problem is called a data sparsity (DS) problem. Another problem is that CVR models trained with samples of clicked impressions infer on samples of all impressions. This problem is called a sample selection bias (SSB) problem. This paper is a summary of the various solutions and current limitations and further directions about solving sample selection bias problem and data sparsity problem.

Original language	English
Title of host publication	5th International Conference on Artificial Intelligence in Information and Communication, ICAIIC 2023
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	232-235
Number of pages	4
ISBN (Electronic)	9781665456456
DOIs	https://doi.org/10.1109/ICAIIC57133.2023.10067098
Publication status	Published - 2023
Event	5th International Conference on Artificial Intelligence in Information and Communication, ICAIIC 2023 - Virtual, Online, Indonesia Duration: 2023 Feb 20 → 2023 Feb 23

Publication series

Name	5th International Conference on Artificial Intelligence in Information and Communication, ICAIIC 2023

Conference

Conference	5th International Conference on Artificial Intelligence in Information and Communication, ICAIIC 2023
Country/Territory	Indonesia
City	Virtual, Online
Period	23/2/20 → 23/2/23

Bibliographical note

Publisher Copyright:
© 2023 IEEE.

Keywords

conversion rate prediction
data sparsity
multi-task learning
recommendation systems
sample selection bias

ASJC Scopus subject areas

Computer Networks and Communications
Information Systems
Signal Processing
Decision Sciences (miscellaneous)
Information Systems and Management
Artificial Intelligence

Access to Document

10.1109/ICAIIC57133.2023.10067098

Cite this

Lee, T., & Seok, J. (2023). Multi Task Learning: A Survey and Future Directions. In 5th International Conference on Artificial Intelligence in Information and Communication, ICAIIC 2023 (pp. 232-235). (5th International Conference on Artificial Intelligence in Information and Communication, ICAIIC 2023). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICAIIC57133.2023.10067098

Multi Task Learning: A Survey and Future Directions. / Lee, Taeho; Seok, Junhee.
5th International Conference on Artificial Intelligence in Information and Communication, ICAIIC 2023. Institute of Electrical and Electronics Engineers Inc., 2023. p. 232-235 (5th International Conference on Artificial Intelligence in Information and Communication, ICAIIC 2023).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Lee, T & Seok, J 2023, Multi Task Learning: A Survey and Future Directions. in 5th International Conference on Artificial Intelligence in Information and Communication, ICAIIC 2023. 5th International Conference on Artificial Intelligence in Information and Communication, ICAIIC 2023, Institute of Electrical and Electronics Engineers Inc., pp. 232-235, 5th International Conference on Artificial Intelligence in Information and Communication, ICAIIC 2023, Virtual, Online, Indonesia, 23/2/20. https://doi.org/10.1109/ICAIIC57133.2023.10067098

Lee T, Seok J. Multi Task Learning: A Survey and Future Directions. In 5th International Conference on Artificial Intelligence in Information and Communication, ICAIIC 2023. Institute of Electrical and Electronics Engineers Inc. 2023. p. 232-235. (5th International Conference on Artificial Intelligence in Information and Communication, ICAIIC 2023). doi: 10.1109/ICAIIC57133.2023.10067098

@inproceedings{543e1a1755904660b8e4897a98461f94,

title = "Multi Task Learning: A Survey and Future Directions",

abstract = "Multi-task learning (MTL) is a problem that must be applied in modern recommendation systems and is just as difficult. In the recent e-commerce advertising market, it is necessary to be able to predict not only the probability of users clicking, but also the probability of conversion and purchase. By predicting multi-task, it is possible to increase the accuracy of each task and optimize advertisements for various goals of advertisers. Traditional conversion rate (CVR) prediction models have difficulty learning because the number of conversions is too small compared to the total number of impressions. This problem is called a data sparsity (DS) problem. Another problem is that CVR models trained with samples of clicked impressions infer on samples of all impressions. This problem is called a sample selection bias (SSB) problem. This paper is a summary of the various solutions and current limitations and further directions about solving sample selection bias problem and data sparsity problem.",

keywords = "conversion rate prediction, data sparsity, multi-task learning, recommendation systems, sample selection bias",

author = "Taeho Lee and Junhee Seok",

note = "Publisher Copyright: {\textcopyright} 2023 IEEE.; 5th International Conference on Artificial Intelligence in Information and Communication, ICAIIC 2023 ; Conference date: 20-02-2023 Through 23-02-2023",

year = "2023",

doi = "10.1109/ICAIIC57133.2023.10067098",

language = "English",

series = "5th International Conference on Artificial Intelligence in Information and Communication, ICAIIC 2023",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "232--235",

booktitle = "5th International Conference on Artificial Intelligence in Information and Communication, ICAIIC 2023",

}

TY - GEN

T1 - Multi Task Learning

T2 - 5th International Conference on Artificial Intelligence in Information and Communication, ICAIIC 2023

AU - Lee, Taeho

AU - Seok, Junhee

PY - 2023

Y1 - 2023

N2 - Multi-task learning (MTL) is a problem that must be applied in modern recommendation systems and is just as difficult. In the recent e-commerce advertising market, it is necessary to be able to predict not only the probability of users clicking, but also the probability of conversion and purchase. By predicting multi-task, it is possible to increase the accuracy of each task and optimize advertisements for various goals of advertisers. Traditional conversion rate (CVR) prediction models have difficulty learning because the number of conversions is too small compared to the total number of impressions. This problem is called a data sparsity (DS) problem. Another problem is that CVR models trained with samples of clicked impressions infer on samples of all impressions. This problem is called a sample selection bias (SSB) problem. This paper is a summary of the various solutions and current limitations and further directions about solving sample selection bias problem and data sparsity problem.

AB - Multi-task learning (MTL) is a problem that must be applied in modern recommendation systems and is just as difficult. In the recent e-commerce advertising market, it is necessary to be able to predict not only the probability of users clicking, but also the probability of conversion and purchase. By predicting multi-task, it is possible to increase the accuracy of each task and optimize advertisements for various goals of advertisers. Traditional conversion rate (CVR) prediction models have difficulty learning because the number of conversions is too small compared to the total number of impressions. This problem is called a data sparsity (DS) problem. Another problem is that CVR models trained with samples of clicked impressions infer on samples of all impressions. This problem is called a sample selection bias (SSB) problem. This paper is a summary of the various solutions and current limitations and further directions about solving sample selection bias problem and data sparsity problem.

KW - conversion rate prediction

KW - data sparsity

KW - multi-task learning

KW - recommendation systems

KW - sample selection bias

UR - http://www.scopus.com/inward/record.url?scp=85151926537&partnerID=8YFLogxK

U2 - 10.1109/ICAIIC57133.2023.10067098

DO - 10.1109/ICAIIC57133.2023.10067098

M3 - Conference contribution

AN - SCOPUS:85151926537

T3 - 5th International Conference on Artificial Intelligence in Information and Communication, ICAIIC 2023

SP - 232

EP - 235

BT - 5th International Conference on Artificial Intelligence in Information and Communication, ICAIIC 2023

PB - Institute of Electrical and Electronics Engineers Inc.

Y2 - 20 February 2023 through 23 February 2023

ER -

Multi Task Learning: A Survey and Future Directions

Abstract

Publication series

Conference

Bibliographical note

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this