DRNet

prior research

• 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).
  • 잠재요인 모형(latent factor models)은 사용자와 아이템의 관계를 모델링하므로 개인화 추천 정확도에 강점을 보임
• FISM Kabbur, S., Ning, X., & Karypis, G. (2013, August). Fism: factored item similarity models for top-n recommender systems. In Proceedings of the 19th ACM SIGKDD international conference on Knowledge discovery and data mining (pp. 659-667).
  • 아이템 기반 협업필터링(user free models)은 목표 아이템(target)과 목표 사용자의 히스토리 아이템(history) 사이의 관계를 모델링하므로 데이터 희소성(data sparcity)에 강건함
• NAIS He, X., He, Z., Song, J., Liu, Z., Jiang, Y. G., & Chua, T. S. (2018). NAIS: Neural attentive item similarity model for recommendation. IEEE Transactions on Knowledge and Data Engineering, 30(12), 2354-2366.
  • 히스토리 아이템을 집계할 때 어텐션 메커니즘(Attention Mechanism)을 활용하여 사용자의 선호 강도에 따라 가중치를 차등 부여하는 것이 균등 부여하는 것보다 사용자의 선호를 효과적으로 대변할 수 있음

idea

• DRNet(Dual Relation Net-work): 사용자-아이템 매칭 함수와 아이템-아이템 매칭 함수를 병렬 학습하는 앙상블 모형

  

notation

• $u=1,2,\cdots,M$: user idx
• $i=1,2,\cdots,N$: item idx
• $\mathbf{X}^{\mathrm{(user)}} \in \mathbb{R}^{M \times M}$: user one-hot matrix
• $\mathbf{X}^{\mathrm{(item)}} \in \mathbb{R}^{N \times N}$: item one-hot matrix
• $\mathbf{Y} \in \mathbb{R}^{M \times N}$: user-item interaction matrix
• $\mathbf{u}_{u} \in \mathbb{R}^{K}$: user id embedding vector @ affection network
• $\mathbf{v}_{i} \in \mathbb{R}^{K}$: item id embedding vector @ affection network
• $\mathbf{p}_{i} \in \mathbb{R}^{K}$: target item id embedding vector @ association network
• $\mathbf{q}_{j} \in \mathbb{R}^{K}$: history item id embedding vector @ association network
• $\mathbf{h} \in \mathbb{R}^{1}$: global query vector @ association network
• $\mathbf{z}_{u,i}$: predictive vector of user $u$ and item $i$
• $\hat{y}_{u,i}$: interaction probability of user $u$ and item $i$

function

• dual-relation networks:

  \[\begin{aligned} \hat{y}_{u,i} &= \sigma\left(\mathbf{W}\cdot\left[\mathbf{z}_{u,i}^{\mathrm{(affect)}}\oplus\mathbf{z}_{u,i}^{\mathrm{(asso)}}\right]\right) \end{aligned}\]

affection network

• index embedding:

  \[\begin{aligned} \mathbf{u}_{u} &=\mathbf{W}\cdot\mathbf{X}_{u*}^{\mathrm{(user)}}\\ \mathbf{v}_{i} &=\mathbf{W}\cdot\mathbf{X}_{i*}^{\mathrm{(item)}} \end{aligned}\]

• matching function learning:

  \[\begin{aligned} \mathbf{z}_{u,i} &= \mathrm{mlp}_{\mathrm{ReLU}}\left(\mathbf{u}_{u}\odot\mathbf{v}_{i}\right) \end{aligned}\]

association network

• index embedding:

  \[\begin{aligned} \mathbf{p}_{i} &=\mathbf{W}\cdot\mathbf{X}_{i*}^{\mathrm{(item)}}\\ \mathbf{q}_{j} &=\mathbf{W}\cdot\mathbf{X}_{j*}^{\mathrm{(item)}} \end{aligned}\]

• global item vector of user $u$:

  \[\begin{aligned} \mathbf{g}_{u} &=\mathrm{att}(\mathbf{h},\mathbf{z}_{u,j}^{\mathrm{(affect)}},\mathbf{q}_{j}),\quad \forall j \in \mathcal{R}_{u}^{+} \setminus \{i\} \end{aligned}\]

• matching function learning:

  \[\begin{aligned} \mathbf{z}_{u,i} &=\mathrm{mlp}_{\mathrm{ReLU}}(\mathbf{g}_{u}\odot\mathbf{p}_{i}) \end{aligned}\]