Collective embedding for neural context-aware recommender systems

Research output: Contribution to book/anthology/report/conference proceedingArticle in proceedingResearchpeer-review

Abstract

Context-aware recommender systems consider contextual features as additional information to predict user's preferences. For example, the recommendations could be based on time, location, or the company of other people. Among the contextual information, time became an important feature because user preferences tend to change over time or be similar in the near future. Researchers have proposed diferent models to incorporate time into their recommender system, however, the current models are not able to capture specifc temporal patterns. To address the limitation observed in previous works, we propose Collective embedding for Neural Context-Aware Recommender Systems (CoNCARS). The proposed solution jointly model the item, user and time embeddings to capture temporal patterns. Then, CoNCARS use the outer product to model the user-item-time correlations between dimensions of the embedding space. The hidden features feed our Convolutional Neural Networks (CNNs) to learn the non-linearities between the diferent features. Finally, we combine the output from our CNNs in the fusion layer and then predict the user's preference score. We conduct extensive experiments on real-world datasets, demonstrating CoNCARS improves the top-N item recommendation task and outperform the state-of-the-art recommendation methods.

Original languageEnglish
Title of host publicationProceedings of the 13th ACM Conference on Recommender Systems, RecSys 2019, Copenhagen, Denmark, September 16-20, 2019.
EditorsToine Bogers, Alan Said, Peter Brusilovsky, Domonkos Tikk
Number of pages9
PublisherAssociation for Computing Machinery
Publication date2019
Pages201-209
ISBN (Print)978-1-4503-6243-6
ISBN (Electronic)9781450362436
DOIs
Publication statusPublished - 2019
EventRecSys 2019: 13th ACM Conference on Recommender Systems - Copenhagen, Denmark, Copenhagen, Denmark
Duration: 16 Sep 201820 Sep 2018
Conference number: 13
http://recsys.acm.org/recsys19

Conference

ConferenceRecSys 2019: 13th ACM Conference on Recommender Systems
Number13
LocationCopenhagen, Denmark
CountryDenmark
CityCopenhagen
Period16/09/201820/09/2018
Internet address

Fingerprint

Recommender systems
Neural networks
Fusion reactions
Industry
Experiments

Keywords

  • Collective Embedding
  • Context-aware Recommender Systems
  • Convolutional Neural Networks
  • Time-aware Recommendations

Cite this

Costa, F. S. D., & Dolog, P. (2019). Collective embedding for neural context-aware recommender systems. In T. Bogers, A. Said, P. Brusilovsky, & D. Tikk (Eds.), Proceedings of the 13th ACM Conference on Recommender Systems, RecSys 2019, Copenhagen, Denmark, September 16-20, 2019. (pp. 201-209). Association for Computing Machinery. https://doi.org/10.1145/3298689.3347028
Costa, Felipe Soares Da ; Dolog, Peter. / Collective embedding for neural context-aware recommender systems. Proceedings of the 13th ACM Conference on Recommender Systems, RecSys 2019, Copenhagen, Denmark, September 16-20, 2019.. editor / Toine Bogers ; Alan Said ; Peter Brusilovsky ; Domonkos Tikk. Association for Computing Machinery, 2019. pp. 201-209
@inproceedings{fe0f119a2dd24e9a975b3f17963904cd,
title = "Collective embedding for neural context-aware recommender systems",
abstract = "Context-aware recommender systems consider contextual features as additional information to predict user's preferences. For example, the recommendations could be based on time, location, or the company of other people. Among the contextual information, time became an important feature because user preferences tend to change over time or be similar in the near future. Researchers have proposed diferent models to incorporate time into their recommender system, however, the current models are not able to capture specifc temporal patterns. To address the limitation observed in previous works, we propose Collective embedding for Neural Context-Aware Recommender Systems (CoNCARS). The proposed solution jointly model the item, user and time embeddings to capture temporal patterns. Then, CoNCARS use the outer product to model the user-item-time correlations between dimensions of the embedding space. The hidden features feed our Convolutional Neural Networks (CNNs) to learn the non-linearities between the diferent features. Finally, we combine the output from our CNNs in the fusion layer and then predict the user's preference score. We conduct extensive experiments on real-world datasets, demonstrating CoNCARS improves the top-N item recommendation task and outperform the state-of-the-art recommendation methods.",
keywords = "Collective Embedding, Context-aware Recommender Systems, Convolutional Neural Networks, Time-aware Recommendations",
author = "Costa, {Felipe Soares Da} and Peter Dolog",
year = "2019",
doi = "10.1145/3298689.3347028",
language = "English",
isbn = "978-1-4503-6243-6",
pages = "201--209",
editor = "Toine Bogers and Alan Said and Peter Brusilovsky and Domonkos Tikk",
booktitle = "Proceedings of the 13th ACM Conference on Recommender Systems, RecSys 2019, Copenhagen, Denmark, September 16-20, 2019.",
publisher = "Association for Computing Machinery",
address = "United States",
note = "RecSys 2019: 13th ACM Conference on Recommender Systems, RecSys 2019 ; Conference date: 16-09-2018 Through 20-09-2018",
url = "http://recsys.acm.org/recsys19",

}

Costa, FSD & Dolog, P 2019, Collective embedding for neural context-aware recommender systems. in T Bogers, A Said, P Brusilovsky & D Tikk (eds), Proceedings of the 13th ACM Conference on Recommender Systems, RecSys 2019, Copenhagen, Denmark, September 16-20, 2019.. Association for Computing Machinery, pp. 201-209, RecSys 2019: 13th ACM Conference on Recommender Systems, Copenhagen, Denmark, 16/09/2018. https://doi.org/10.1145/3298689.3347028

Collective embedding for neural context-aware recommender systems. / Costa, Felipe Soares Da; Dolog, Peter.

Proceedings of the 13th ACM Conference on Recommender Systems, RecSys 2019, Copenhagen, Denmark, September 16-20, 2019.. ed. / Toine Bogers; Alan Said; Peter Brusilovsky; Domonkos Tikk. Association for Computing Machinery, 2019. p. 201-209.

Research output: Contribution to book/anthology/report/conference proceedingArticle in proceedingResearchpeer-review

TY - GEN

T1 - Collective embedding for neural context-aware recommender systems

AU - Costa, Felipe Soares Da

AU - Dolog, Peter

N1 - Conference code: 13

PY - 2019

Y1 - 2019

N2 - Context-aware recommender systems consider contextual features as additional information to predict user's preferences. For example, the recommendations could be based on time, location, or the company of other people. Among the contextual information, time became an important feature because user preferences tend to change over time or be similar in the near future. Researchers have proposed diferent models to incorporate time into their recommender system, however, the current models are not able to capture specifc temporal patterns. To address the limitation observed in previous works, we propose Collective embedding for Neural Context-Aware Recommender Systems (CoNCARS). The proposed solution jointly model the item, user and time embeddings to capture temporal patterns. Then, CoNCARS use the outer product to model the user-item-time correlations between dimensions of the embedding space. The hidden features feed our Convolutional Neural Networks (CNNs) to learn the non-linearities between the diferent features. Finally, we combine the output from our CNNs in the fusion layer and then predict the user's preference score. We conduct extensive experiments on real-world datasets, demonstrating CoNCARS improves the top-N item recommendation task and outperform the state-of-the-art recommendation methods.

AB - Context-aware recommender systems consider contextual features as additional information to predict user's preferences. For example, the recommendations could be based on time, location, or the company of other people. Among the contextual information, time became an important feature because user preferences tend to change over time or be similar in the near future. Researchers have proposed diferent models to incorporate time into their recommender system, however, the current models are not able to capture specifc temporal patterns. To address the limitation observed in previous works, we propose Collective embedding for Neural Context-Aware Recommender Systems (CoNCARS). The proposed solution jointly model the item, user and time embeddings to capture temporal patterns. Then, CoNCARS use the outer product to model the user-item-time correlations between dimensions of the embedding space. The hidden features feed our Convolutional Neural Networks (CNNs) to learn the non-linearities between the diferent features. Finally, we combine the output from our CNNs in the fusion layer and then predict the user's preference score. We conduct extensive experiments on real-world datasets, demonstrating CoNCARS improves the top-N item recommendation task and outperform the state-of-the-art recommendation methods.

KW - Collective Embedding

KW - Context-aware Recommender Systems

KW - Convolutional Neural Networks

KW - Time-aware Recommendations

U2 - 10.1145/3298689.3347028

DO - 10.1145/3298689.3347028

M3 - Article in proceeding

SN - 978-1-4503-6243-6

SP - 201

EP - 209

BT - Proceedings of the 13th ACM Conference on Recommender Systems, RecSys 2019, Copenhagen, Denmark, September 16-20, 2019.

A2 - Bogers, Toine

A2 - Said, Alan

A2 - Brusilovsky, Peter

A2 - Tikk, Domonkos

PB - Association for Computing Machinery

T2 - RecSys 2019: 13th ACM Conference on Recommender Systems

Y2 - 16 September 2018 through 20 September 2018

ER -

Costa FSD, Dolog P. Collective embedding for neural context-aware recommender systems. In Bogers T, Said A, Brusilovsky P, Tikk D, editors, Proceedings of the 13th ACM Conference on Recommender Systems, RecSys 2019, Copenhagen, Denmark, September 16-20, 2019.. Association for Computing Machinery. 2019. p. 201-209 https://doi.org/10.1145/3298689.3347028