Abstract
Online recommender systems aim to provide timely recommended results by constantly updating the model with new interactions. However, existing methods require sufficient personal data and fail to accurately support online recommendations for cold-start users. Although the state-of-the-art method adopts meta-learning to solve this problem, it requires to recall previously seen data for model update, which is impractical due to linearly increasing memory over time. In this paper, we propose a memory-efficient online meta-recommender MEOM that can avoid the explicit use of historical data while achieving high accuracy. The recommender adopts MAML as a meta-learner and particularly adapts it to online scenarios with effective regularization. Specifically, an online regularization method is designed to summarize the time-varying model and historical task gradients, such that overall model optimization direction can be acquired to parameterize a meaningful regularizer as a penalty for next round model update. The regularizer is then utilized to guide model updates with prior knowledge in a memory-efficient and accurate manner. Besides, to avoid task-overfitting, an adaptive learning rate strategy is adopted to control model adaptation by more suitable learning rates in dual levels. Experimental results on two real-world datasets show that our method can significantly reduce memory consumption while keeping accuracy performance.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
Our code is available at: https://anonymous.4open.science/r/MEOM-master-4239,
- 2.
- 3.
References
Acar, D.A.E., Zhu, R., Saligrama, V.: Memory efficient online meta learning. In: ICML 2021, pp. 32ā42. PMLR (2021)
Al-Ghossein, M., Abdessalem, T., Barre, A.: A survey on stream-based recommender systems. ACM Comput. Surv. (CSUR) 54(5), 1ā36 (2021)
Chen, F., Luo, M., Dong, Z., etĀ al.: Federated meta-learning with fast convergence and efficient communication. arXiv preprint arXiv:1802.07876 (2018)
Cheng, H.T., Koc, L., Harmsen, J., etĀ al.: Wide & deep learning for recommender systems. In: DLRS 2016, pp. 7ā10 (2016)
Dong, M., Yuan, F., Yao, L., etĀ al.: Mamo: memory-augmented meta-optimization for cold-start recommendation. In: SIGKDD 2020, pp. 688ā697 (2020)
Du, Z., Wang, X., Yang, H., etĀ al.: Sequential scenario-specific meta learner for online recommendation. In: SIGKDD 2019, pp. 2895ā2904 (2019)
Finn, C., Abbeel, P., Levine, S.: Model-agnostic meta-learning for fast adaptation of deep networks. In: ICML 2017, pp. 1126ā1135. PMLR (2017)
Kim, M., Song, H., Shin, Y., etĀ al.: Meta-learning for online update of recommender systems. arXiv preprint arXiv:2203.10354 (2022)
Lee, H., Im, J., Jang, S., etĀ al.: MeLU: meta-learned user preference estimator for cold-start recommendation. In: SIGKDD 2019, pp. 1073ā1082 (2019)
Li, J., Jing, M., Lu, K., etĀ al.: From zero-shot learning to cold-start recommendation. In: AAAI 2019, vol.Ā 33, pp. 4189ā4196 (2019)
Li, Y., Xu, J., et al.: ATLRec: an attentional adversarial transfer learning network for cross-domain recommendation. J. Comput. Sci. Technol. 35(4), 794ā808 (2020)
Lv, Y., Xu, J., Zhou, R., Fang, J., Liu, C.: SSRGAN: a generative adversarial network for streaming sequential recommendation. In: Jensen, C.S., et al. (eds.) DASFAA 2021. LNCS, vol. 12683, pp. 36ā52. Springer, Cham (2021). https://doi.org/10.1007/978-3-030-73200-4_3
Song, J., Xu, J., Zhou, R., etĀ al.: CBML: a cluster-based meta-learning model for session-based recommendation. In: CIKM, pp. 1713ā1722 (2021)
Sun, H., Xu, J., Zheng, K., etĀ al.: MFNP: a meta-optimized model for few-shot next POI recommendation. In: Zhou, Z. (ed.) IJCAI 2021, pp. 3017ā3023 (2021)
Sun, H., Xu, J., Zhou, R., et al.: HOPE: a hybrid deep neural model for out-of-town next POI recommendation. WWW 24(5), 1749ā1768 (2021)
Sun, X., Shi, T., Gao, X., etĀ al.: FORM: follow the online regularized meta-leader for cold-start recommendation. In: SIGIR 2021, pp. 1177ā1186 (2021)
Wang, L., Jin, B., Huang, Z., etĀ al.: Preference-adaptive meta-learning for cold-start recommendation. In: IJCAI 2021, pp. 1607ā1614 (2021)
Wang, Q., Yin, H., Hu, Z., etĀ al.: Neural memory streaming recommender networks with adversarial training. In: SIGKDD 2018, pp. 2467ā2475 (2018)
Wang, W., Yin, H., etĀ al.: Streaming ranking based recommender systems. In: SIGIR 2018, pp. 525ā534 (2018)
Xu, J., Song, J., Sang, Y., et al.: CDAML: a cluster-based domain adaptive meta-learning model for cross domain recommendation. WWW 26(3), 989ā1003 (2023)
Yu, B., Li, X., Fang, J., et al.: Memory-augmented meta-learning framework for session-based target behavior recommendation. WWW 26(1), 233ā251 (2023)
Yu, R., Gong, Y., He, X., etĀ al.: Personalized adaptive meta learning for cold-start user preference prediction. In: AAAI 2021, vol.Ā 35, pp. 10772ā10780 (2021)
Zhang, Y., Feng, etĀ al.: How to retrain recommender system? A sequential meta-learning method. In: SIGIR 2020, pp. 1479ā1488 (2020)
Zhou, G., Zhu, X., Song, C., etĀ al.: Deep interest network for click-through rate prediction. In: SIGKDD 2018, pp. 1059ā1068 (2018)
Zhu, F., Wang, Y., Chen, C., etĀ al.: Cross-domain recommendation: challenges, progress, and prospects. arXiv preprint arXiv:2103.01696 (2021)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
Ā© 2024 The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
About this paper
Cite this paper
Luo, Y., Zhang, R. (2024). MEOM: Memory-Efficient Online Meta-recommender forĀ Cold-Start Recommendation. In: Song, X., Feng, R., Chen, Y., Li, J., Min, G. (eds) Web and Big Data. APWeb-WAIM 2023. Lecture Notes in Computer Science, vol 14332. Springer, Singapore. https://doi.org/10.1007/978-981-97-2390-4_23
Download citation
DOI: https://doi.org/10.1007/978-981-97-2390-4_23
Published:
Publisher Name: Springer, Singapore
Print ISBN: 978-981-97-2389-8
Online ISBN: 978-981-97-2390-4
eBook Packages: Computer ScienceComputer Science (R0)