TY - CHAP

T1 - A Unified Contextual Bandit Framework for Long- and Short-Term Recommendations

AU - Tavakol, Maryam

AU - Brefeld, Ulf

N1 - Conference code: 27

PY - 2017/12/30

Y1 - 2017/12/30

N2 - We present a unified contextual bandit framework for recommendation problems that is able to capture long- and short-term interests of users. The model is devised in dual space and the derivation is consequentially carried out using Fenchel-Legrende conjugates and thus leverages to a wide range of tasks and settings.We detail two instantiations for regression and classification scenarios and obtain well-known algorithms for these special cases. The resulting general and unified framework allows for quickly adapting contextual bandits to different applications at-hand. The empirical study demonstrates that the proposed long- and short-term framework outperforms both, short-term and long-term models on data. Moreover, a tweak of the combined model proves beneficial in cold start problems.

AB - We present a unified contextual bandit framework for recommendation problems that is able to capture long- and short-term interests of users. The model is devised in dual space and the derivation is consequentially carried out using Fenchel-Legrende conjugates and thus leverages to a wide range of tasks and settings.We detail two instantiations for regression and classification scenarios and obtain well-known algorithms for these special cases. The resulting general and unified framework allows for quickly adapting contextual bandits to different applications at-hand. The empirical study demonstrates that the proposed long- and short-term framework outperforms both, short-term and long-term models on data. Moreover, a tweak of the combined model proves beneficial in cold start problems.

KW - Informatics

KW - Machine learning

KW - Recommender systems

KW - Contextual bandits

KW - Recommendation

KW - Dual optimization

KW - Personalization

UR - http://ecmlpkdd2017.ijs.si/papers/paperID239.pdf

UR - http://www.scopus.com/inward/record.url?scp=85040222425&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-71246-8_17

DO - 10.1007/978-3-319-71246-8_17

M3 - Article in conference proceedings

SN - 978-3-319-71245-1

VL - 2

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 269

EP - 284

BT - Machine Learning and Knowledge Discovery in Databases

A2 - Ceci, Michelangelo

A2 - Hollmen, Jaakko

A2 - Todorovski, Ljupco

A2 - Vens, Celine

A2 - Dzeroski, Saso

PB - Springer

T2 - THE EUROPEAN CONFERENCE ON MACHINE LEARNING & PRINCIPLES AND PRACTICE OF KNOWLEDGE DISCOVERY IN DATABASES 2017

Y2 - 18 September 2017 through 22 September 2017

ER -