Post

J-NCF

Chen, W., Cai, F., Chen, H., & Rijke, M. D.
(2019).
Joint neural collaborative filtering for recommender systems.
ACM Transactions on Information Systems (TOIS), 37(4), 1-30.

prior research

  • DMF Xue, H. J., Dai, X., Zhang, J., Huang, S., & Chen, J. (2017, August). Deep matrix factorization models for recommender systems. In IJCAI (Vol. 17, pp. 3203-3209).
    • 표현 학습(Feature Learning)은 사용자와 아이템의 관계가 쌍선형(Bilinear)이라는 단순 가정 하에 엔티티들의 의미론적(Basis) 유사도를 반영하는 의미 체계(Linear Space, Latent Space)를 학습하는 데 강점이 있음
  • NCF He, X., Liao, L., Zhang, H., Nie, L., Hu, X., & Chua, T. S. (2017, April). Neural collaborative filtering. In Proceedings of the 26th international conference on world wide web (pp. 173-182).
    • 매칭 함수 학습(Interaction Learning)은 사용자와 아이템의 관계를 단순 가정으로 사전 제약하지 않고 데이터로부터 학습하여 데이터에 적합한 관계 구조를 모사하는 데 강점이 있음
  • DeepCF Deng, Z. H., Huang, L., Wang, C. D., Lai, J. H., & Yu, P. S. (2019, July). Deepcf: A unified framework of representation learning and matching function learning in recommender system. In Proceedings of the AAAI conference on artificial intelligence (Vol. 33, No. 01, pp. 61-68).
    • 해당 모형(DeepCF)은 표현 학습과 매칭 함수 학습을 별개의 과정으로 구분하므로 상호 보완되지 못하여 상호작용 모델링이 제한되거나 엔티티 표현력이 제한됨

idea

  • J-NCF(Joint Neural Collaborative Filtering): 표현 학습 모듈(Representation Learning Networks)과 매칭 학습 모듈(Matching Function Learning Networks)을 직렬 결합하여 통합 훈련(Joint Training)하는 모형

    01

notation

  • $u=1,2,\cdots,M$: user idx
  • $i=1,2,\cdots,N$: item idx
  • $\mathbf{Y} \in \mathbb{R}^{M \times N}$: user-item interaction matrix
  • $\mathbf{u}_{u} \in \mathbb{R}^{K}$: user latent factor vector
  • $\mathbf{v}_{i} \in \mathbb{R}^{K}$: item latent factor vector
  • $\mathbf{z}_{u,i}$: predictive vector of user $u$ and item $i$
  • $\hat{y}_{u,i}$: interaction probability of user $u$ and item $i$

function

  • user latent factor vector representation learning:

    \[\begin{aligned} \mathbf{u}_{u} &=\mathrm{mlp}_{\mathrm{ReLU}}(\mathbf{Y}_{u*}) \end{aligned}\]
  • item latent factor vector representation learning:

    \[\begin{aligned} \mathbf{v}_{i} &=\mathrm{mlp}_{\mathrm{ReLU}}(\mathbf{Y}_{*i}) \end{aligned}\]
  • matching function learning:

    \[\begin{aligned} \mathbf{z}_{u,i} &=\mathrm{mlp}_{\mathrm{ReLU}}(\mathbf{u}_{u} \oplus\mathbf{v}_{i}) \end{aligned}\]
  • predict interaction probability of user $u$ and item $i$:

    \[\begin{aligned} \hat{y}_{u,i} &= \sigma(\mathbf{w} \cdot \mathbf{z}_{u,i}) \end{aligned}\]
This post is licensed under CC BY 4.0 by the author.