IJCAI19 RS presentation整理

Coupling Everything: A Universal Guideline for Building State-of-The-Art Recommender Systems

Challenges in Collaborated Filtering:

Data Sparsity
Cold Start
Scalability

Challenges in Content-based Filtering

Limited Content Analysis
Over-specialization

以上问题的来源是因为数据的缺乏，所以要做信息的coupling。
A. Coupling on users:

Social RS: 用户相互作用
Group RS：集体智能决策

B. Coupling on items:

Cross Domain RS: 领域交叉
Session-based RS: 时序耦合

C. Coupling on implicit interaction

Context-aware RS: 上下文依赖
多目标评分
吸引：主观关注

Songlei Jian, Liang Hu, Longbing Cao & Kai Lu. AAAI 2018. Metric-based Auto-Instructor for Learning Mixed Data Representation

Tensor Factorization

Network Embedding

data representation for RS

Coupling User:

Tang, J., Hu, X., & Liu, H. (2013). Social recommendation: a review. Social Network Analysis and Mining, 3(4), 1113-1133.

Social RS: user mutual influence

Ma, H., Yang, H., Lyu, M. R., & King, I. (2008, October). Sorec: social recommendation using probabilistic matrix factorization. In Proceedings of the 17th ACM conference on Information and knowledge management (pp. 931-940). ACM.

Jamali, M., & Ester, M. (2010, September). A matrix factorization technique with trust propagation for recommendation in social networks. In Proceedings of the fourth ACM conference on Recommender systems (pp. 135-142). ACM.

Ma, H., Zhou, D., Liu, C., Lyu, M. R., & King, I. (2011, February). Recommender systems with social regularization. In Proceedings of the fourth ACM international conference on Web search and data mining (pp. 287-296). ACM.

Wang, X., He, X., Nie, L., & Chua, T. S. (2017). Item Silk Road: Recommending Items from Information Domains to Social Users. arXiv preprint arXiv:1706.03205.

Social RS 三个发展方向：
Network embedding
Memory mechanism
Dynamic model

Group RS:
Profile Aggregation
GPA/IPA
Masthoff, J. (2015). Group recommender systems: aggregation, satisfaction and group attributes. In Recommender Systems Handbook (pp. 743-776).

Least misery strategy
Average strategy

Lu, Q., Yang, D., Cheng, T., Zhang, W., Yu, Y. Informative Household Recommendation with Feature-based Matrix Factorization. In CAMRa2011, 2011.

Hu, L., Cao, J., Xu, G., Cao, L., Gu, Z., and Cao, W. Deep modeling of group preferences for group-based recommendation. In Twenty-Eighth AAAI Conference on Artificial Intelligence, 2014.

Cao, D., He, X. , et al. Attentive group recommendation. In SIGIR 2018, 2018.

Group RS 三个发展方向
先看问题：

Lack of group feedback data
Dynamic representation given a group of users
Contextual information

Direction:
context -aware group recommendation

Coupling items

Cross Domain RS:
naive MF for cross domain: 就是User-item表堆起来
用一个表说明了item之间的统一性，但其实这是不对的，但是其实这会隐藏不同item之间的不同性质。

Pan, W., Xiang, E. W., Liu, N. N., & Yang, Q. (2010, July). Transfer Learning in Collaborative Filtering for Sparsity Reduction. In AAAI (Vol. 10, pp. 230-235).

Hu, L., Cao, L., Cao, J., Gu, Z., Xu, G., & Yang, D. (2016). Learning Informative Priors from Heterogeneous Domains to Improve Recommendation in Cold-Start User Domains. ACM Transactions on Information Systems (TOIS), 35(2), 13.

Elkahky, A.M., Song, Y., and He, X. A multi-view deep learning approach for cross domain user modeling in recommendation systems. In Proceedings of the 24th International Conference on World Wide Web, 278-288, 2015.

Kim, T., Cha, M., Kim, H., Lee, J., & Kim, J. (2017). Learning to discover cross-domain relations with generative adversarial networks. arXiv preprint arXiv:1703.05192.

DiscoGAN

问题和方向
信息迁移
用不重叠的物品和用户联合进行多领域学习

Session-based RS:
为什么要对session进行建模：
因为对历史账户建模会重复推荐相似物品

First order dependency modeling
Rendle, S., Freudenthaler, C., and Schmidt-Thieme , L. (2010, August). Factorizing Personalized Markov Chains for Next-Basket Recommendation. WWW2010.

Higher order dependency modeling
GRU4rec
Hidasi, B., Karatzoglou,A., Baltrunas, L., and Tikk, D. (2016, May). Session-based Recommendations with Recurrent Neural Networks. ICLR2016.

Loosely ordered sequence is session
Hu, L., Cao, L., Wang, S., Cao, J., Gu, Z., Xu, G., and Wang, J. Diversifying Personalized Recommendation with User-session Context. IJCAI2017

Hu, L., Cao, L., Wang, S., Cao, J., Gu, Z., Xu, G., and Wang, J. Diversifying Personalized Recommendation with User-session Context. IJCAI2017

NTEM

Wang, S., Hu, L., & Cao, L. (2017, September). Perceiving the Next Choice with Comprehensive Transaction Embeddings for Online Recommendation. In Joint European Conference on Machine Learning and Knowledge Discovery in Databases(pp. 285-302). Springer, Cham.

Tang J., Wang K. (2018, February). Personalized Top-N Sequential Recommendation via Convolutional Sequence Embedding. WSDM2018.

Tang J., Belletti F. , etc. (2019, May). Towards Neural Mixture Recommender for Long Range Dependent User Sequences. WWW2019.

SRS

Wang S., Cao L. , etc. (2019, February). A Survey on Session-based Recommender Systems. Arxive2019.

Multi-modal RS
multimodal learning

Oramas, S., Nieto, O., Sordo, M., & Serra, X. (2017). A deep multimodal approach for cold-start music recommendation. arXiv preprint arXiv:1706.09739.

Lynch, C., Aryafar, K., and Attenberg, J. Images Don't Lie: Transferring Deep Visual Semantic Features to Large-Scale Multimodal Learning to Rank. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 541-548, 2016.

多模态的意义在于一个模态可以帮助另一个模态去区分一些东西
Image information can help disentangle different listings considered similar by a text model

如何解决不同模态下噪音的问题：
Multi-view learning over heterogeneous data sources
• Extract consistent information from review, description, item images, user videos
• GAN-based models to generate multiple types of samples • Text, images, videos

多评价标准的推荐系统学习
shilling attack detection

Shilling attack detection for recommender systems based on credibility of group users and rating time series

Hu, L., Cao, L., Cao, J., Gu, Z., Xu, G., and Wang, J. Improving the Quality of Recommendations for Users and Items in the Tail of Distribution. ACM Trans. Inf. Syst., 2017

这块不是很熟，先mark下
问题和方向：
How to integrate the impacts from multiple criteria?
• Different users may pay attention to different objectives
• The importance of objectives are often dependent on the context
• Modeling with game theory to find equilibria over multiple criteria
• Applying multi-objective optimization methods in RS • Multiple-criteria decision analysis
• Multidisciplinary design optimization

Comprehensive couplings in all

Adomavicius, G. and Tuzhilin, A., 2015. Context-aware recommender systems. In Recommender systems handbook (pp. 191-226). Springer US.

Context is any factor (observable or not observable) leading to user behavior
Karatzoglou, A., et al. Multiverse recommendation: n-dimensional tensor factorization for context-aware collaborative filtering. In Recsys, 79-86, 2010.

Cheng, H. T., Koc, L., Harmsen, J., Shaked, T., Chandra, T., Aradhye, H., ... & Anil, R. (2016, September). Wide & deep learning for recommender systems. In Proceedings of the 1st Workshop on Deep Learning for Recommender Systems (pp. 7-10). ACM.

Li Y., Cao L. , etc. (2017, August). Mining fashion outfit composition using an end-to-end deep learning approach on set data. IEEE Transactions on Multimedia

Cucurull, G., Taslakian, P. and Vazquez, D., Context-Aware Visual Compatibility Prediction. In CVPR 2019

Kang, W., Fang, C., Wang, Z., and McAuley, J. Visually-Aware Fashion Recommendation and Design with Generative Image Models. In 2017 IEEE ICDM

Hu, L., Jian, S., Cao, L., Chen. Q. Interpretable Recommendation via Attraction Modeling: Learning Multilevel Attractiveness over Multimodal Movie Contents. IJCAI2018

Modeling Attraction

问题和未来的方向
open issues and directions

Hu, L., Jian, S., Cao, L., Gu, Z., Chen, Q., Amirbekyan, A. HERS: Modeling Influential Contexts with Heterogeneous Relations for Sparse and Cold-start Recommendation. In AAAI-19

Jian, S., Hu, L., Cao, L., Lu, K., and Gao, H. Evolutionarily Learning Multi-aspect Interactions and Influences from Network Structure and Node Content. In AAAI-19

conclusion