Back

Deep Learning Based Recommender System

Shuai ZhangLina YaoAixin SunYi Tay

article2017ACM Computing Surveys1,345 citations

Provides a comprehensive taxonomy of deep learning recommender systems, synthesizing state-of-the-art architectures and outlining critical open directions for future research.

Listen

Online platforms face severe information overload, making personalized recommendation essential for user engagement and business revenue. Major industry leaders rely heavily on recommender systems—driving 80% of movies watched on Netflix and 60% of video clicks on YouTube. While traditional recommendation methods such as linear matrix factorization have served as industry workhorses, they struggle to capture complex non-linear user-item relationships and cannot easily ingest unstructured multimedia data like text, images, and audio. The article comprehensively evaluates the landscape of deep learning based recommender systems to categorize existing models, assess their strengths and trade-offs, and establish directions for future research and deployment.

The analysis systematically reviews over 100 recent academic studies and industrial implementations from major venues. The authors establish a two-dimensional taxonomy that classifies techniques by their foundational neural building blocks—including multilayer perceptrons, autoencoders, convolutional neural networks, recurrent neural networks, restricted Boltzmann machines, neural autoregressive estimators, attention mechanisms, adversarial networks, and reinforcement learning—as well as composite hybrid architectures that combine these techniques for specialized tasks.

The review yields five key findings. First, deep neural networks deliver substantial performance gains over linear baselines by capturing complex non-linear interaction patterns and automatically learning feature representations without labor-intensive manual feature engineering. Second, deep architectures excel at unifying heterogeneous, multi-modal data sources (such as product review texts, images, and audio signals) alongside historical interaction data within end-to-end trainable pipelines. Third, specialized architectures solve critical operational constraints: recurrent and convolutional models effectively capture sequential patterns and temporal dynamics for session-based settings where user identifiers are absent, while reinforcement learning enables real-time adaptation to streaming user feedback. Fourth, neural attention mechanisms provide a dual benefit of boosting recommendation accuracy while mitigating the traditional "black-box" limitation by offering interpretability into why specific items are recommended. Finally, empirical evidence demonstrates that deep collaborative models typically achieve optimal performance at relatively shallow depths of three to four layers, beyond which performance plateaus.

These findings indicate that adopting deep learning frameworks can reduce engineering costs related to manual feature extraction while significantly improving recommendation accuracy and user retention. However, practitioners must account for the computational demands of large-scale deep architectures and recognize that simpler neighborhood or linear methods remain competitive in basic interaction-only scenarios. Organizations transitioning to deep recommender architectures should selectively deploy models tailored to their primary data modality—such as recurrent or attention networks for sequential web sessions and convolutional networks for rich media—while leveraging attention weights to maintain explainability for end users and system auditors.

Looking forward, the article highlights critical open research challenges, noting that the field urgently requires standardized, blinded benchmark datasets and rigorous evaluation protocols to ensure consistent, reliable progress. Future research and pilot implementations should prioritize model compression and knowledge distillation to scale inference to massive production workloads, develop robust cross-domain transfer learning methods, and explore multi-task learning architectures that simultaneously improve recommendation quality and generate actionable textual explanations.

Nanyang Technological UniversityUniversity of New South Walescomputer visiondeep learninginformation overloadsdeep recommender systemsDeep Collaborative FilteringRecurrent Recommender Network
arXiv: 1707.07435cheungdaven/DeepRec
Cover for Deep Learning Based Recommender System

Abstract

With the ever-growing volume of online information, recommender systems have been an effective strategy to overcome such information overload. The utility of recommender systems cannot be overstated, given its widespread adoption in many web applications, along with its potential impact to ameliorate many problems related to over-choice. In recent years, deep learning has garnered considerable interest in many research fields such as computer vision and natural language processing, owing not only to stellar performance but also the attractive property of learning feature representations from scratch. The influence of deep learning is also pervasive, recently demonstrating its effectiveness when applied to information retrieval and recommender systems research. Evidently, the field of deep learning in recommender system is flourishing. This article aims to provide a comprehensive review of recent research efforts on deep learning based recommender systems. More concretely, we provide and devise a taxonomy of deep learning based recommendation models, along with providing a comprehensive summary of the state-of-the-art. Finally, we expand on current trends and provide new perspectives pertaining to this new exciting development of the field.

Topics

computer visiondeep learninginformation overloadsrecommendationtaxonomyweb applicationcollaborative filteringrecommendation systemsrecommendation algorithmspersonalized recommendation

Keywords

deep recommender systemsDeep Collaborative FilteringRecurrent Recommender NetworkDeep Factorization MachineBoltzmann Machine Based RecommendationNeural Network Matrix Factorizationcontent-based recommender systemsequential recommendationsession-based recommendationcontext-aware recommendation

Table of Contents

  • 1 Introduction
  • 2 Overview of Recommender Systems and Deep Learning
  • 2.1 Recommender Systems
  • 2.2 Deep Learning Techniques
  • 2.3 Why Deep Neural Networks for Recommendation?
  • 2.4 On Potential Limitations
  • 3 Deep Learning Based Recommendation: State-of-the-art
  • 3.1 Categories of deep learning based recommendation models
  • 3.2 Multilayer Perceptron based Recommendation
  • 3.3 Autoencoder based Recommendation
  • 3.4 Convolutional Neural Networks based Recommendation
  • 3.5 Recurrent Neural Networks based Recommendation
  • 3.6 Restricted Boltzmann Machine based Recommendation
  • 3.7 Neural Attention based Recommendation
  • 3.8 Neural AutoRegressive based Recommendation
  • 3.9 Deep Reinforcement Learning for Recommendation
  • 3.10 Adversarial Network based Recommendation
  • 3.11 Deep Hybrid Models for Recommendation
  • 4 Future Research Directions and Open Issues
  • 4.1 Joint Representation Learning from User and Item Content Information
  • 4.2 Explainable Recommendation with Deep Learning
  • 4.3 Going Deeper for Recommendation
  • 4.4 Machine Reasoning for Recommendation
  • 4.5 Cross Domain Recommendation with Deep Neural Networks
  • 4.6 Deep Multi-Task Learning for Recommendation
  • 4.7 Scalability of Deep Neural Networks for Recommendation
  • 4.8 The Field Needs Better, More Unified and Harder Evaluation
  • 5 Conclusion
  • References

Knowls

  1. Knowl 1 — Taxonomy of Deep Learning Based Recommender Systems

    model/method

    Deep learning based recommender systems are categorized into two primary paradigms:

    1. Recommendation with Neural Building Blocks: Architectures that rely on a single primary neural network class, grouped into:

      • Multilayer Perceptron (MLP): Models non-linear user-item interactions and feature embeddings (e.g., NCF, DeepFM, Wide & Deep, DSSM).
      • Autoencoders (AE): Employs bottleneck representations for feature learning or directly reconstructs partially observed interaction vectors (e.g., AutoRec, CFN, CDAE, Multi-VAE).
      • Convolutional Neural Networks (CNNs): Extracts hierarchical visual, textual, and unstructured features from auxiliary data, or models high-order interaction matrices and graph structures.
      • Recurrent Neural Networks (RNNs): Models sequential dynamics, temporal evolution of user preferences, and session-based interactions (e.g., GRU4Rec, RRN).
      • Restricted Boltzmann Machines (RBM): Employs two-layer energy models for collaborative filtering.
      • Neural Autoregressive Distribution Estimation (NADE): Performs tractable generative probability estimation over rating vectors (e.g., CF-NADE).
      • Neural Attention Models (AM): Employs soft attention or co-attention mechanisms to dynamically select informative items/words and improve model interpretability.
      • Adversarial Networks (AN): Unifies generative and discriminative retrieval via minimax optimization (e.g., IRGAN).
      • Deep Reinforcement Learning (DRL): Treats recommendation as a dynamic, interactive trial-and-error decision process with adaptive reward functions.

    2. Recommendation with Deep Hybrid Models: Architectures that combine two or more neural paradigms (e.g., CNN + Autoencoder in CKE, CNN + RNN in neural citation networks, or RNN + DRL for dynamic treatment recommendations) to capture complementary data characteristics simultaneously.

  2. Knowl 2 — Neural Collaborative Filtering Framework

    model/method

    Neural Collaborative Filtering (NCF) replaces the traditional inner product interaction function of matrix factorization with a dual-path feed-forward neural network architecture.

    Let suusers_u^{\text{user}} and siitems_i^{\text{item}} denote the input feature representations (such as one-hot identifiers or side information vectors) for user uu and item ii, and let U∈RM×kU \in \mathbb{R}^{M \times k} and V∈RN×kV \in \mathbb{R}^{N \times k} denote the user and item latent embedding matrices with dimension kk. The predicted interaction score r^ui\hat{r}_{ui} is formulated as:

    r^ui=f(UTsuuser,VTsiitem∣U,V,θ)\hat{r}_{ui} = f\left(U^T s_u^{\text{user}}, V^T s_i^{\text{item}} \mid U, V, \theta\right)

    where f(⋅)f(\cdot) is a multilayer perceptron parameterized by θ\theta.

    When optimized for implicit feedback with binary labels rui∈{0,1}r_{ui} \in \{0, 1\}, the network is trained using binary cross-entropy loss over the set of observed interactions O\mathcal{O} and sampled unobserved instances O−\mathcal{O}^-:

    L=−∑(u,i)∈O∪O−(ruilog⁡r^ui+(1−rui)log⁡(1−r^ui))\mathcal{L} = -\sum_{(u,i) \in \mathcal{O} \cup \mathcal{O}^-} \left( r_{ui} \log \hat{r}_{ui} + (1 - r_{ui}) \log(1 - \hat{r}_{ui}) \right)

    Traditional matrix factorization is a linear special case of this framework, and linear factorizations can be fused with non-linear MLP pathways into generalized collaborative filtering networks.

  3. Knowl 3 — Deep Factorization Machine Architecture

    model/method

    The Deep Factorization Machine (DeepFM) integrates a Factorization Machine (FM) engine with a Multilayer Perceptron (MLP) into an end-to-end architecture to simultaneously model low-order and high-order feature interactions without requiring manual feature engineering.

    For an input feature vector xx composed of mm fields (representing user ID, item ID, and contextual attributes), the model consists of two components sharing the same underlying dense feature embeddings:

    1. The FM component, which uses addition and inner products to capture first-order linear features and second-order pairwise feature interactions, yielding scalar output yFM(x)y_{FM}(x).
    2. The MLP component, which feeds the concatenated field embeddings through multiple fully connected layers with non-linear activations to capture high-order feature interactions, yielding scalar output yMLP(x)y_{MLP}(x).

    The combined prediction score r^ui\hat{r}_{ui} is computed as:

    r^ui=σ(yFM(x)+yMLP(x))\hat{r}_{ui} = \sigma\left( y_{FM}(x) + y_{MLP}(x) \right)

    where σ(z)=11+e−z\sigma(z) = \frac{1}{1 + e^{-z}} is the sigmoid activation function.

  4. Knowl 4 — Wide and Deep Learning Architecture

    model/method

    The Wide & Deep learning model combines a generalized linear model with a deep feed-forward neural network to achieve both memorization of historical correlation patterns and generalization to unseen feature combinations.

    • The wide component computes:

      ywide=WwideT{x,ϕ(x)}+by_{\text{wide}} = W_{\text{wide}}^T \{x, \phi(x)\} + b

      where xx is the raw input feature vector, ϕ(x)\phi(x) represents cross-product transformations that capture domain-specific feature interactions, and Wwide,bW_{\text{wide}}, b are model parameters.

    • The deep component computes feed-forward layer activations:

      α(l+1)=f(Wdeep(l)α(l)+b(l))\alpha^{(l+1)} = f\left(W_{\text{deep}}^{(l)} \alpha^{(l)} + b^{(l)}\right)

      where ll denotes layer depth, f(⋅)f(\cdot) is the activation function, and Wdeep(l),b(l)W_{\text{deep}}^{(l)}, b^{(l)} are weights and biases.

    The combined prediction probability for binary rating labels r^ui∈{0,1}\hat{r}_{ui} \in \{0, 1\} is:

    P(r^ui=1∣x)=σ(WwideT{x,ϕ(x)}+WdeepTa(lf)+bias)P(\hat{r}_{ui} = 1 \mid x) = \sigma\left(W_{\text{wide}}^T \{x, \phi(x)\} + W_{\text{deep}}^T a^{(l_f)} + \text{bias}\right)

    where a(lf)a^{(l_f)} is the final layer activation of the deep component, and σ(⋅)\sigma(\cdot) is the sigmoid function.

  5. Knowl 5 — AutoRec Framework for Collaborative Rating Reconstruction

    model/method

    AutoRec applies autoencoders to collaborative filtering by reconstructing partially observed rating vectors directly. It has two formulations: Item-based AutoRec (I-AutoRec) and User-based AutoRec (U-AutoRec).

    In I-AutoRec, each item i∈{1,…,N}i \in \{1, \dots, N\} is represented by its partially observed rating vector r(i)=(r1i,…,rMi)T∈RMr^{(i)} = (r_{1i}, \dots, r_{Mi})^T \in \mathbb{R}^M across MM users. The reconstruction function h(r(i);θ)h(r^{(i)}; \theta) is defined as:

    h(r(i);θ)=f(W⋅g(V⋅r(i)+μ)+b)h(r^{(i)}; \theta) = f\left( W \cdot g(V \cdot r^{(i)} + \mu) + b \right)

    where V∈Rk×MV \in \mathbb{R}^{k \times M} and W∈RM×kW \in \mathbb{R}^{M \times k} are projection and reconstruction parameter matrices for latent dimension kk, μ∈Rk\mu \in \mathbb{R}^k and b∈RMb \in \mathbb{R}^M are bias vectors, f(⋅)f(\cdot) and g(⋅)g(\cdot) are activation functions, and θ={W,V,μ,b}\theta = \{W, V, \mu, b\}.

    The optimization objective minimizes reconstruction error over only the observed ratings set O\mathcal{O}:

    min⁡θ∑i=1N∥r(i)−h(r(i);θ)∥O2+λ⋅reg\min_\theta \sum_{i=1}^N \| r^{(i)} - h(r^{(i)}; \theta) \|_{\mathcal{O}}^2 + \lambda \cdot \text{reg}

    where ∥⋅∥O2\|\cdot\|_{\mathcal{O}}^2 computes squared loss strictly over observed entries, and reg\text{reg} represents weight regularization scaled by hyperparameter λ\lambda.

  6. Knowl 6 — Collaborative Denoising Autoencoder for Top-N Recommendation

    model/method

    Collaborative Denoising Auto-Encoder (CDAE) adapts denoising autoencoders to top-NN recommendation with implicit feedback. The input is a user's partially observed binary preference vector rpref(u)∈{0,1}Nr_{\text{pref}}^{(u)} \in \{0, 1\}^N over NN items, which is corrupted by noise (e.g., Gaussian or masking noise) to yield r~pref(u)∼p(r~pref(u)∣rpref(u))\tilde{r}_{\text{pref}}^{(u)} \sim p(\tilde{r}_{\text{pref}}^{(u)} \mid r_{\text{pref}}^{(u)}).

    CDAE introduces a dedicated user-specific latent embedding node Vu∈RKV_u \in \mathbb{R}^K into the reconstruction function:

    h(r~pref(u))=f(W2⋅g(W1r~pref(u)+Vu+b1)+b2)h(\tilde{r}_{\text{pref}}^{(u)}) = f\left( W_2 \cdot g(W_1 \tilde{r}_{\text{pref}}^{(u)} + V_u + b_1) + b_2 \right)

    where W1∈RK×NW_1 \in \mathbb{R}^{K \times N} and W2∈RN×KW_2 \in \mathbb{R}^{N \times K} are weight matrices, b1∈RKb_1 \in \mathbb{R}^K and b2∈RNb_2 \in \mathbb{R}^N are bias terms, and f(⋅),g(⋅)f(\cdot), g(\cdot) are activation functions.

    The parameters are trained by minimizing the reconstruction loss over all MM users with negative sampling:

    min⁡W1,W2,V,b1,b21M∑u=1MEp(r~pref(u)∣rpref(u))[ℓ(rpref(u),h(r~pref(u)))]+λ⋅reg\min_{W_1, W_2, V, b_1, b_2} \frac{1}{M} \sum_{u=1}^M \mathbb{E}_{p(\tilde{r}_{\text{pref}}^{(u)} \mid r_{\text{pref}}^{(u)})} \left[ \ell\left( r_{\text{pref}}^{(u)}, h(\tilde{r}_{\text{pref}}^{(u)}) \right) \right] + \lambda \cdot \text{reg}

    where ℓ(⋅)\ell(\cdot) is cross-entropy or squared error loss, and reg\text{reg} is regularizing weight decay.

  7. Knowl 7 — Collaborative Deep Learning Bayesian Framework

    model/method

    Collaborative Deep Learning (CDL) is a hierarchical Bayesian framework that tightly couples a perception component—a probabilistic Stacked Denoising Autoencoder (SDAE)—with a task-specific recommendation component (Probabilistic Matrix Factorization, PMF) to balance content representations and interaction history.

    The Bayesian generative process is formulated as follows:

    1. For each layer l∈{1,…,L}l \in \{1, \dots, L\} of the SDAE:
      • Weight columns Wl,∗n∼N(0,λw−1IDl)W_{l,*n} \sim \mathcal{N}(0, \lambda_w^{-1} I_{D_l}) and bias vector bl∼N(0,λw−1IDl)b_l \sim \mathcal{N}(0, \lambda_w^{-1} I_{D_l}).
      • Layer representations Xl,i∗∼N(σ(Xl−1,i∗Wl+bl),λs−1IDl)X_{l,i*} \sim \mathcal{N}(\sigma(X_{l-1,i*} W_l + b_l), \lambda_s^{-1} I_{D_l}).
    2. For each item ii:
      • Clean input Xc,i∗∼N(XL,i∗,λn−1IIi)X_{c,i*} \sim \mathcal{N}(X_{L,i*}, \lambda_n^{-1} I_{I_i}).
      • Item offset vector ϵi∼N(0,λv−1ID)\epsilon_i \sim \mathcal{N}(0, \lambda_v^{-1} I_D), with the final item latent vector set to Vi=ϵi+XL/2,i∗TV_i = \epsilon_i + X_{L/2,i*}^T, where XL/2,i∗X_{L/2,i*} is the bottleneck layer of the autoencoder.
    3. For each user uu, latent factor Uu∼N(0,λu−1ID)U_u \sim \mathcal{N}(0, \lambda_u^{-1} I_D).
    4. For each user-item interaction (u,i)(u, i), rating rui∼N(UuTVi,Cui−1)r_{ui} \sim \mathcal{N}(U_u^T V_i, C_{ui}^{-1}), where CuiC_{ui} is an observation confidence parameter.

    The framework is optimized using coordinate descent / expectation-maximization updates.

  8. Knowl 8 — Unified Deep Collaborative Filtering Framework

    model/method

    The Deep Collaborative Filtering (DCF) framework unifies arbitrary deep neural feature representation architectures with collaborative filtering objectives. Given an interaction matrix RR, user auxiliary information XX, and item auxiliary information YY, the generalized optimization problem is defined as:

    min⁡U,Vℓ(R,U,V)+β(∥U∥F2+∥V∥F2)+γL(X,U)+δL(Y,V)\min_{U, V} \ell(R, U, V) + \beta \left( \|U\|_F^2 + \|V\|_F^2 \right) + \gamma \mathcal{L}(X, U) + \delta \mathcal{L}(Y, V)

    where:

    • ℓ(R,U,V)\ell(R, U, V) is the collaborative filtering loss function operating over user latent factors UU and item latent factors VV.
    • ∥U∥F2+∥V∥F2\|U\|_F^2 + \|V\|_F^2 is the Frobenius norm regularization on latent factors.
    • L(X,U)\mathcal{L}(X, U) and L(Y,V)\mathcal{L}(Y, V) represent differentiable loss terms acting as structural hinges that connect deep neural feature learning models (such as marginalized autoencoders, CNNs, or RNNs) to the collaborative latent factor spaces UU and VV.
    • β,γ,δ\beta, \gamma, \delta are trade-off hyperparameters balancing the collaborative filtering loss and the side-information representation losses.

  9. Knowl 9 — Collaborative Filtering with Neural Autoregressive Distribution Estimation

    model/method

    Collaborative Filtering with Neural Autoregressive Distribution Estimation (CF-NADE) provides a tractable probability estimator for user ratings, serving as a closed-form alternative to Restricted Boltzmann Machines (RBM-CF).

    For a user rating vector rr, CF-NADE computes the exact joint probability of observed ratings using the probability chain rule across an ordering o=(o1,…,oD)o = (o_1, \dots, o_D) of the DD items rated by the user:

    p(r)=∏i=1Dp(rmoi∣rmo<i)p(r) = \prod_{i=1}^D p(r_{m_{o_i}} \mid r_{m_{o_{<i}}})

    where moim_{o_i} is the index of the ii-th rated item, rmoir_{m_{o_i}} is the assigned rating score, and rmo<ir_{m_{o_{<i}}} denotes the preceding subset of item ratings in the ordering oo. Conditional probabilities are computed via a feedforward architecture with tied hidden parameter matrices across autoregressive steps, enabling linear computational complexity with respect to the number of observed ratings.

  10. Knowl 10 — Information Retrieval Generative Adversarial Network Framework

    model/method

    IRGAN applies Generative Adversarial Networks (GANs) to information retrieval and item recommendation through a minimax game between a generative retrieval model pθ(d∣q,r)p_\theta(d \mid q, r) and a discriminative retrieval model fϕ(q,d)f_\phi(q, d):

    • The generative model pθ(d∣q,r)p_\theta(d \mid q, r) approximates the true underlying preference distribution ptrue(d∣q,r)p_{\text{true}}(d \mid q, r) by generating candidate items dd given user query or profile qq to confuse the discriminator.
    • The discriminative model fϕ(q,d)f_\phi(q, d) predicts whether a given pair (q,d)(q, d) is a true ground-truth interaction or an artificial candidate generated by pθp_\theta.

    The unified minimax objective is:

    JG∗,D∗=min⁡θmax⁡ϕ∑n=1N(Ed∼ptrue(d∣qn,r)[log⁡D(d∣qn)]+Ed∼pθ(d∣qn,r)[log⁡(1−D(d∣qn))])J^{G^*, D^*} = \min_\theta \max_\phi \sum_{n=1}^N \left( \mathbb{E}_{d \sim p_{\text{true}}(d \mid q_n, r)} \left[ \log D(d \mid q_n) \right] + \mathbb{E}_{d \sim p_\theta(d \mid q_n, r)} \left[ \log\left( 1 - D(d \mid q_n) \right) \right] \right)

    where D(d∣q)=σ(fϕ(q,d))D(d \mid q) = \sigma(f_\phi(q, d)), and parameters θ\theta and ϕ\phi are updated alternately via policy-gradient reinforcement learning and gradient descent.

  11. Knowl 11 — Session-Based Recommendation via Recurrent Neural Networks

    model/method

    Session-based recommendation using Gated Recurrent Units (GRU4Rec) models sequential user click actions in anonymous web sessions where persistent user IDs are unavailable. The input at each session step is a 1-of-NN one-hot encoding of the active item across NN total items, and the network outputs the likelihood of each candidate item being the next interaction.

    To optimize ranking without full softmax computation over large item vocabularies, the network is trained using session-parallel mini-batches and a pairwise ranking loss termed the TOP1 loss:

    Ls=1S∑j=1S(σ(r^sj−r^si)+σ(r^sj2))\mathcal{L}_s = \frac{1}{S} \sum_{j=1}^S \left( \sigma(\hat{r}_{sj} - \hat{r}_{si}) + \sigma(\hat{r}_{sj}^2) \right)

    where SS is the sample size, r^sj\hat{r}_{sj} and r^si\hat{r}_{si} are the predicted scores for the ground-truth positive item jj and a sampled negative item ii at session step ss, and σ(⋅)\sigma(\cdot) is the logistic sigmoid function. The second term penalizes high absolute positive prediction scores to regularize learning.

  12. Knowl 12 — Identified Open Research Challenges in Deep Recommender Systems

    limitation

    Deep learning based recommender systems exhibit several systematic challenges and structural limitations:

    1. Performance Saturation with Depth: In interaction-only collaborative filtering settings (neural CF), model accuracy typically plateaus at three to four layers, with deeper feed-forward architectures failing to yield further gains without specialized auxiliary losses, residual connections, or layer-wise adaptive learning rates.
    2. Evaluation Inconsistency and Baselines: Recommender systems research suffers from non-standardized dataset splits, arbitrary negative sampling protocols, and unblinded test evaluations. Well-tuned simple neighborhood baselines can frequently match or outperform complex neural models when evaluated under identical conditions.
    3. Black-box Explainability and Multi-step Reasoning: Although attention weights identify influential words or items, current deep architectures lack true multi-step relational reasoning over heterogeneous modalities (e.g., reviews, knowledge graph meta-paths, and visual features) required to generate coherent conversational explanations.
    4. Scalability and Real-time Inference: Deploying billion-parameter multi-modal models to high-throughput streaming environments requires advances in incremental online training, embedding compression, and knowledge distillation from teacher networks to compact inference models.

Coverage note — Specific single-application models (such as niche dance background or makeup recommenders) and standard textbook definitions of generic neural building blocks (basic MLPs, CNNs, LSTMs) were omitted in favor of core recommendation architectures, generalized formulations, taxonomy, and survey-level synthesis.

References

  1. 1.Gediminas Adomavicius and Alexander Tuzhilin. 2005. Toward the next generation of recommender systems: A survey of the state-of-the-art and possible extensions. IEEE transactions on knowledge and data engineering 17, 6 (2005), 734–749.
  2. 2.Taleb Alashkar, Songyao Jiang, Shuyang Wang, and Yun Fu. 2017. Examples-Rules Guided Deep Neural Network for Makeup Recommendation. In AAAI. 941–947.
  3. 3.Jimmy Ba, Volodymyr Mnih, and Koray Kavukcuoglu. 2014. Multiple object recognition with visual attention. arXiv preprint arXiv:1412.7755 (2014).
  4. 4.Bing Bai, Yushun Fan, Wei Tan, and Jia Zhang. 2017. DLTSR: A Deep Learning Framework for Recommendation of Long-tail Web Services. IEEE Transactions on Services Computing (2017).
  5. 5.Trapit Bansal, David Belanger, and Andrew McCallum. 2016. Ask the gru: Multi-task learning for deep text recommendations. In Proceedings of the 10th ACM Conference on Recommender Systems. 107–114.
  6. 6.Rianne van den Berg, Thomas N Kipf, and Max Welling. 2017. Graph convolutional matrix completion. arXiv preprint arXiv:1706.02263 (2017).
  7. 7.Basiliyos Tilahun Betru, Charles Awono Onana, and Bernabe Batchakui. 2017. Deep Learning Methods on Recommender System: A Survey of State-of-the-art. International Journal of Computer Applications 162, 10 (Mar 2017).
  8. 8.Robin Burke. 2002. Hybrid recommender systems: Survey and experiments. User modeling and user-adapted interaction 12, 4 (2002), 331–370.
  9. 9.Xiaoyan Cai, Junwei Han, and Libin Yang. 2018. Generative Adversarial Network Based Heterogeneous Bibliographic Network Representation for Personalized Citation Recommendation. In AAAI.
  10. 10.S. Cao, N. Yang, and Z. Liu. 2017. Online news recommender based on stacked auto-encoder. In ICIS. 721–726.
  11. 11.Rose Catherine and William Cohen. 2017. Transnets: Learning to transform for recommendation. In Recsys. 288–296.
  12. 12.Cheng Chen, Xiangwu Meng, Zhenghua Xu, and Thomas Lukasiewicz. 2017. Location-Aware Personalized News Recommendation With Deep Semantic Analysis. IEEE Access 5 (2017), 1624–1638.
  13. 13.Cen Chen, Peilin Zhao, Longfei Li, Jun Zhou, Xiaolong Li, and Minghui Qiu. 2017. Locally Connected Deep Learning Framework for Industrial-scale Recommender Systems. In WWW.
  14. 14.Jingyuan Chen, Hanwang Zhang, Xiangnan He, Liqiang Nie, Wei Liu, and Tat-Seng Chua. 2017. Attentive Collaborative Filtering: Multimedia Recommendation with Item- and Component-Level Attention. (2017).
  15. 15.Minmin Chen, Zhixiang Xu, Kilian Weinberger, and Fei Sha. 2012. Marginalized denoising autoencoders for domain adaptation. arXiv preprint arXiv:1206.4683 (2012).
  16. 16.Shi-Yong Chen, Yang Yu, Qing Da, Jun Tan, Hai-Kuan Huang, and Hai-Hong Tang. 2018. Stabilizing reinforcement learning in dynamic environment with application to online recommendation. In SIGKDD. 1187–1196.
  17. 17.Xu Chen, Yongfeng Zhang, Qingyao Ai, Hongteng Xu, Junchi Yan, and Zheng Qin. 2017. Personalized Key Frame Recommendation. In SIGIR.
  18. 18.Xu Chen, Yongfeng Zhang, Hongteng Xu, Yixin Cao, Zheng Qin, and Hongyuan Zha. 2018. Visually Explainable Recommendation. arXiv preprint arXiv:1801.10288 (2018).
  19. 19.Yifan Chen and Maarten de Rijke. 2018. A Collective Variational Autoencoder for Top-N Recommendation with Side Information. arXiv preprint arXiv:1807.05730 (2018).
  20. 20.Heng-Tze Cheng, Levent Koc, Jeremiah Harmsen, Tal Shaked, Tushar Chandra, Hrishi Aradhye, Glen Anderson, Greg Corrado, Wei Chai, Mustafa Ispir, and others. 2016. Wide & deep learning for recommender systems. In Recsys. 7–10.
  21. 21.Sungwoon Choi, Heonseok Ha, Uiwon Hwang, Chanju Kim, Jung-Woo Ha, and Sungroh Yoon. 2018. Reinforcement Learning based Recommender System using Biclustering Technique. arXiv preprint arXiv:1801.05532 (2018).
  22. 22.Jan Chorowski, Dzmitry Bahdanau, Kyunghyun Cho, and Yoshua Bengio. 2014. End-to-end continuous speech recognition using attention-based recurrent NN: first results. arXiv preprint arXiv:1412.1602 (2014).
  23. 23.Jan K Chorowski, Dzmitry Bahdanau, Dmitriy Serdyuk, Kyunghyun Cho, and Yoshua Bengio. 2015. Attention-based models for speech recognition. In Advances in Neural Information Processing Systems. 577–585.
  24. 24.Konstantina Christakopoulou, Alex Beutel, Rui Li, Sagar Jain, and Ed H Chi. 2018. Q&R: A Two-Stage Approach toward Interactive Recommendation. In SIGKDD. 139–148.
  25. 25.Wei-Ta Chu and Ya-Lun Tsai. 2017. A hybrid recommendation system considering visual information for predicting favorite restaurants. WWWJ (2017), 1–19.
  26. 26.Ronan Collobert and Jason Weston. 2008. A unified architecture for natural language processing: Deep neural networks with multitask learning. In Proceedings of the 25th international conference on Machine learning. 160–167.
  27. 27.Paul Covington, Jay Adams, and Emre Sargin. 2016. Deep neural networks for youtube recommendations. In Recsys. 191–198.
  28. 28.Hanjun Dai, Yichen Wang, Rakshit Trivedi, and Le Song. 2016. Deep coevolutionary network: Embedding user and item features for recommendation. arXiv preprint. arXiv preprint arXiv:1609.03675 (2016).
  29. 29.Hanjun Dai, Yichen Wang, Rakshit Trivedi, and Le Song. 2016. Recurrent coevolutionary latent feature processes for continuous-time recommendation. In Recsys. 29–34.
  30. 30.James Davidson, Benjamin Liebald, Junning Liu, Palash Nandy, Taylor Van Vleet, Ullas Gargi, Sujoy Gupta, Yu He, Mike Lambert, Blake Livingston, and Dasarathi Sampath. 2010. The YouTube Video Recommendation System. In Recsys.
  31. 31.Li Deng, Dong Yu, and others. 2014. Deep learning: methods and applications. Foundations and Trends® in Signal Processing 7, 3–4 (2014), 197–387.
  32. 32.Shuiguang Deng, Longtao Huang, Guandong Xu, Xindong Wu, and Zhaohui Wu. 2017. On deep learning for trust-aware recommendations in social networks. IEEE transactions on neural networks and learning systems 28, 5 (2017), 1164–1177.
  33. 33.Robin Devooght and Hugues Bersini. 2016. Collaborative filtering with recurrent neural networks. arXiv preprint arXiv:1608.07400 (2016).
  34. 34.Xin Dong, Lei Yu, Zhonghuo Wu, Yuxia Sun, Lingfeng Yuan, and Fangxi Zhang. 2017. A Hybrid Collaborative Filtering Model with Deep Structure for Recommender Systems. In AAAI. 1309–1315.
  35. 35.Tim Donkers, Benedikt Loepp, and Jürgen Ziegler. 2017. Sequential user-based recurrent neural network recommendations. In Recsys. 152–160.
  36. 36.Chao Du, Chongxuan Li, Yin Zheng, Jun Zhu, and Bo Zhang. 2016. Collaborative Filtering with User-Item Co-Autoregressive Models. arXiv preprint arXiv:1612.07146 (2016).
  37. 37.Gintare Karolina Dziugaite and Daniel M Roy. 2015. Neural network matrix factorization. arXiv preprint arXiv:1511.06443 (2015).
  38. 38.Travis Ebesu and Yi Fang. 2017. Neural Citation Network for Context-Aware Citation Recommendation. (2017).
  39. 39.Ali Mamdouh Elkahky, Yang Song, and Xiaodong He. 2015. A multi-view deep learning approach for cross domain user modeling in recommendation systems. In WWW. 278–288.
  40. 40.Ignacio Fernández-Tobías, Iván Cantador, Marius Kaminskas, and Francesco Ricci. 2012. Cross-domain recommender systems: A survey of the state of the art. In Spanish Conference on Information Retrieval. 24.
  41. 41.Jianfeng Gao, Li Deng, Michael Gamon, Xiaodong He, and Patrick Pantel. 2014. Modeling interestingness with deep neural networks. (June 13 2014). US Patent App. 14/304,863.
  42. 42.Kostadin Georgiev and Preslav Nakov. 2013. A non-iid framework for collaborative filtering with restricted boltzmann machines. In ICML. 1148–1156.
  43. 43.Carlos A Gomez-Uribe and Neil Hunt. 2016. The netflix recommender system: Algorithms, business value, and innovation. TMIS 6, 4 (2016), 13.
  44. 44.Yuyun Gong and Qi Zhang. 2016. Hashtag Recommendation Using Attention-Based Convolutional Neural Network.. In IJCAI. 2782–2788.
  45. 45.Ian Goodfellow, Yoshua Bengio, and Aaron Courville. 2016. Deep Learning. MIT Press. http://www.deeplearningbook.org.
  46. 46.Ian Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. 2014. Generative adversarial nets. In NIPS. 2672–2680.
  47. 47.Huifeng Guo, Ruiming Tang, Yunming Ye, Zhenguo Li, and Xiuqiang He. 2017. DeepFM: A Factorization-Machine based Neural Network for CTR Prediction. In IJCAI. 2782–2788.
  48. 48.Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. 2016. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition. 770–778.
  49. 49.Ruining He and Julian McAuley. 2016. Ups and downs: Modeling the visual evolution of fashion trends with one-class collaborative filtering. In WWW. 507–517.
  50. 50.Ruining He and Julian McAuley. 2016. VBPR: Visual Bayesian Personalized Ranking from Implicit Feedback. In AAAI. 144–150.
  51. 51.Xiangnan He, Xiaoyu Du, Xiang Wang, Feng Tian, Jinhui Tang, and Tat-Seng Chua. 2018. Outer Product-based Neural Collaborative Filtering. (2018).
  52. 52.Xiangnan He, Zhankui He, Xiaoyu Du, and Tat-Seng Chua. 2018. Adversarial Personalized Ranking for Recommendation. In SIGIR. 355–364.
  53. 53.Xiangnan He, Lizi Liao, Hanwang Zhang, Liqiang Nie, Xia Hu, and Tat-Seng Chua. 2017. Neural collaborative filtering. In WWW. 173–182.
  54. 54.Xiangnan He and Chua Tat-Seng. 2017. Neural Factorization Machines for Sparse Predictive Analytics. (2017).
  55. 55.Balázs Hidasi and Alexandros Karatzoglou. 2017. Recurrent neural networks with top-k gains for session-based recommendations. arXiv preprint arXiv:1706.03847 (2017).
  56. 56.Balázs Hidasi, Alexandros Karatzoglou, Linas Baltrunas, and Domonkos Tikk. 2015. Session-based recommendations with recurrent neural networks. International Conference on Learning Representations (2015).
  57. 57.Balázs Hidasi, Massimo Quadrana, Alexandros Karatzoglou, and Domonkos Tikk. 2016. Parallel recurrent neural network architectures for feature-rich session-based recommendations. In Recsys. 241–248.
  58. 58.Kurt Hornik. 1991. Approximation capabilities of multilayer feedforward networks. Neural networks 4, 2 (1991), 251–257.
  59. 59.Kurt Hornik, Maxwell Stinchcombe, and Halbert White. 1989. Multilayer feedforward networks are universal approximators. Neural networks 2, 5 (1989), 359–366.
  60. 60.Cheng-Kang Hsieh, Longqi Yang, Yin Cui, Tsung-Yi Lin, Serge Belongie, and Deborah Estrin. 2017. Collaborative metric learning. In WWW. 193–201.
  61. 61.Cheng-Kang Hsieh, Longqi Yang, Honghao Wei, Mor Naaman, and Deborah Estrin. 2016. Immersive recommendation: News and event recommendations using personal digital traces. In WWW. 51–62.
  62. 62.Binbin Hu, Chuan Shi, Wayne Xin Zhao, and Philip S Yu. 2018. Leveraging Meta-path based Context for Top-N Recommendation with A Neural Co-Attention Model. In SIGKDD. 1531–1540.
  63. 63.Yifan Hu, Yehuda Koren, and Chris Volinsky. 2008. Collaborative Filtering for Implicit Feedback Datasets. In ICDM.
  64. 64.Gao Huang, Zhuang Liu, Laurens Van Der Maaten, and Kilian Q Weinberger. 2017. Densely Connected Convolutional Networks.. In CVPR, Vol. 1. 3.
  65. 65.Po-Sen Huang, Xiaodong He, Jianfeng Gao, Li Deng, Alex Acero, and Larry Heck. 2013. Learning deep structured semantic models for web search using clickthrough data. In CIKM. 2333–2338.
  66. 66.Wenyi Huang, Zhaohui Wu, Liang Chen, Prasenjit Mitra, and C Lee Giles. 2015. A Neural Probabilistic Model for Context Based Citation Recommendation. In AAAI. 2404–2410.
  67. 67.Drew A Hudson and Christopher D Manning. 2018. Compositional attention networks for machine reasoning. arXiv preprint arXiv:1803.03067 (2018).
  68. 68.Dietmar Jannach and Malte Ludewig. 2017. When Recurrent Neural Networks Meet the Neighborhood for Session-Based Recommendation. In Recsys.
  69. 69.Dietmar Jannach, Markus Zanker, Alexander Felfernig, and Gerhard Friedrich. 2010. Recommender systems: an introduction.
  70. 70.Yogesh Jhamb, Travis Ebesu, and Yi Fang. 2018. Attentive Contextual Denoising Autoencoder for Recommendation. (2018).
  71. 71.X. Jia, X. Li, K. Li, V. Gopalakrishnan, G. Xun, and A. Zhang. 2016. Collaborative restricted Boltzmann machine for social event recommendation. In ASONAM. 402–405.
  72. 72.Xiaowei Jia, Aosen Wang, Xiaoyi Li, Guangxu Xun, Wenyao Xu, and Aidong Zhang. 2015. Multi-modal learning for video recommendation based on mobile application usage. In 2015 IEEE International Conference on Big Data (Big Data). 837–842.
  73. 73.How Jing and Alexander J Smola. 2017. Neural survival recommender. In WSDM. 515–524.
  74. 74.Muhammad Murad Khan, Roliana Ibrahim, and Imran Ghani. 2017. Cross Domain Recommender Systems: A Systematic Literature Review. ACM Comput. Surv. 50, 3 (June 2017).
  75. 75.Donghyun Kim, Chanyoung Park, Jinoh Oh, Sungyoung Lee, and Hwanjo Yu. 2016. Convolutional matrix factorization for document context-aware recommendation. In Recsys. 233–240.
  76. 76.Donghyun Kim, Chanyoung Park, Jinoh Oh, and Hwanjo Yu. 2017. Deep Hybrid Recommender Systems via Exploiting Document Context and Statistics of Items. Information Sciences (2017).
  77. 77.Thomas N Kipf and Max Welling. 2016. Semi-supervised classification with graph convolutional networks. arXiv preprint arXiv:1609.02907 (2016).
  78. 78.Young-Jun Ko, Lucas Maystre, and Matthias Grossglauser. 2016. Collaborative recurrent neural networks for dynamic recommender systems. In Asian Conference on Machine Learning. 366–381.
  79. 79.Yehuda Koren. 2008. Factorization meets the neighborhood: a multifaceted collaborative filtering model. In SIGKDD. 426–434.
  80. 80.Yehuda Koren. 2010. Collaborative filtering with temporal dynamics. Commun. ACM 53, 4 (2010), 89–97.
  81. 81.Hugo Larochelle and Iain Murray. 2011. The neural autoregressive distribution estimator. In Proceedings of the Fourteenth International Conference on Artificial Intelligence and Statistics. 29–37.
  82. 82.Hanbit Lee, Yeonchan Ahn, Haejun Lee, Seungdo Ha, and Sang-goo Lee. 2016. Quote Recommendation in Dialogue using Deep Neural Network. In SIGIR. 957–960.
  83. 83.Joonseok Lee, Sami Abu-El-Haija, Balakrishnan Varadarajan, and Apostol Paul Natsev. 2018. Collaborative Deep Metric Learning for Video Understanding. (2018).
  84. 84.Chenyi Lei, Dong Liu, Weiping Li, Zheng-Jun Zha, and Houqiang Li. 2016. Comparative Deep Learning of Hybrid Representations for Image Recommendations. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2545–2553.
  85. 85.Jure Leskovec. 2015. New Directions in Recommender Systems. In WSDM.
  86. 86.Lihong Li, Wei Chu, John Langford, and Robert E Schapire. 2010. A contextual-bandit approach to personalized news article recommendation. In Proceedings of the 19th international conference on World wide web. 661–670.
  87. 87.Piji Li, Zihao Wang, Zhaochun Ren, Lidong Bing, and Wai Lam. 2017. Neural Rating Regression with Abstractive Tips Generation for Recommendation. (2017).
  88. 88.Sheng Li, Jaya Kawale, and Yun Fu. 2015. Deep collaborative filtering via marginalized denoising auto-encoder. In CIKM. 811–820.
  89. 89.Xiaopeng Li and James She. 2017. Collaborative Variational Autoencoder for Recommender Systems. In SIGKDD.
  90. 90.Yang Li, Ting Liu, Jing Jiang, and Liang Zhang. 2016. Hashtag recommendation with topical attention-based LSTM. In COLING.
  91. 91.Zhi Li, Hongke Zhao, Qi Liu, Zhenya Huang, Tao Mei, and Enhong Chen. 2018. Learning from History and Present: Next-item Recommendation via Discriminatively Exploiting User Behaviors. In SIGKDD. 1734–1743.
  92. 92.Jianxun Lian, Fuzheng Zhang, Xing Xie, and Guangzhong Sun. 2017. CCCFNet: A Content-Boosted Collaborative Filtering Neural Network for Cross Domain Recommender Systems. In WWW. 817–818.
  93. 93.Jianxun Lian, Xiaohuan Zhou, Fuzheng Zhang, Zhongxia Chen, Xing Xie, and Guangzhong Sun. 2018. xDeepFM: Combining Explicit and Implicit Feature Interactions for Recommender Systems. arXiv preprint arXiv:1803.05170 (2018).
  94. 94.Dawen Liang, Rahul G Krishnan, Matthew D Hoffman, and Tony Jebara. 2018. Variational Autoencoders for Collaborative Filtering. arXiv preprint arXiv:1802.05814 (2018).
  95. 95.Dawen Liang, Minshu Zhan, and Daniel PW Ellis. 2015. Content-Aware Collaborative Music Recommendation Using Pre-trained Neural Networks.. In ISMIR. 295–301.
  96. 96.Yankai Lin, Zhiyuan Liu, Maosong Sun, Yang Liu, and Xuan Zhu. 2015. Learning Entity and Relation Embeddings for Knowledge Graph Completion. In AAAI. 2181–2187.
  97. 97.Juntao Liu and Caihua Wu. 2017. Deep Learning Based Recommendation: A Survey.
  98. 98.Qiang Liu, Shu Wu, and Liang Wang. 2017. DeepStyle: Learning User Preferences for Visual Recommendation. (2017).
  99. 99.Qiao Liu, Yifu Zeng, Refuoe Mokhosi, and Haibin Zhang. 2018. STAMP: Short-Term Attention/Memory Priority Model for Session-based Recommendation. In SIGKDD. 1831–1839.
  100. 100.Xiaomeng Liu, Yuanxin Ouyang, Wenge Rong, and Zhang Xiong. 2015. Item Category Aware Conditional Restricted Boltzmann Machine Based Recommendation. In International Conference on Neural Information Processing. 609–616.
  101. 101.Pablo Loyola, Chen Liu, and Yu Hirate. 2017. Modeling User Session and Intent with an Attention-based Encoder-Decoder Architecture. In Recsys. 147–151.
  102. 102.Pablo Loyola, Chen Liu, and Yu Hirate. 2017. Modeling User Session and Intent with an Attention-based Encoder-Decoder Architecture. In Recsys (RecSys ’17).
  103. 103.Malte Ludewig and Dietmar Jannach. 2018. Evaluation of Session-based Recommendation Algorithms. CoRR abs/1803.09587 (2018).
  104. 104.Minh-Thang Luong, Hieu Pham, and Christopher D Manning. 2015. Effective approaches to attention-based neural machine translation. arXiv preprint arXiv:1508.04025 (2015).
  105. 105.Julian McAuley, Christopher Targett, Qinfeng Shi, and Anton Van Den Hengel. 2015. Image-based recommendations on styles and substitutes. In SIGIR. 43–52.
  106. 106.Volodymyr Mnih, Koray Kavukcuoglu, David Silver, Andrei A Rusu, Joel Veness, Marc G Bellemare, Alex Graves, Martin Riedmiller, Andreas K Fidjeland, Georg Ostrovski, and others. 2015. Human-level control through deep reinforcement learning. Nature 518, 7540 (2015), 529.
  107. 107.Isshu Munemasa, Yuta Tomomatsu, Kunioki Hayashi, and Tomohiro Takagi. 2018. Deep Reinforcement Learning for Recommender Systems. (2018).
  108. 108.Cataldo Musto, Claudio Greco, Alessandro Suglia, and Giovanni Semeraro. 2016. Ask Me Any Rating: A Content-based Recommender System based on Recurrent Neural Networks. In IIR.
  109. 109.Maryam M Najafabadi, Flavio Villanustre, Taghi M Khoshgoftaar, Naeem Seliya, Randall Wald, and Edin Muharemagic. 2015. Deep learning applications and challenges in big data analytics. Journal of Big Data 2, 1 (2015), 1.
  110. 110.Hanh T. H. Nguyen, Martin Wistuba, Josif Grabocka, Lucas Rego Drumond, and Lars Schmidt-Thieme. 2017. Personalized Deep Learning for Tag Recommendation.
  111. 111.Xia Ning and George Karypis. 2010. Multi-task learning for recommender system. In Proceedings of 2nd Asian Conference on Machine Learning. 269–284.
  112. 112.Wei Niu, James Caverlee, and Haokai Lu. 2018. Neural Personalized Ranking for Image Recommendation. In Proceedings of the Eleventh ACM International Conference on Web Search and Data Mining. 423–431.
  113. 113.Shumpei Okura, Yukihiro Tagami, Shingo Ono, and Akira Tajima. 2017. Embedding-based News Recommendation for Millions of Users. In SIGKDD.
  114. 114.Yuanxin Ouyang, Wenqi Liu, Wenge Rong, and Zhang Xiong. 2014. Autoencoder-based collaborative filtering. In International Conference on Neural Information Processing. 284–291.
  115. 115.Weike Pan, Evan Wei Xiang, Nathan Nan Liu, and Qiang Yang. 2010. Transfer Learning in Collaborative Filtering for Sparsity Reduction. In AAAI, Vol. 10. 230–235.
  116. 116.Yiteng Pana, Fazhi Hea, and Haiping Yua. 2017. Trust-aware Collaborative Denoising Auto-Encoder for Top-N Recommendation. arXiv preprint arXiv:1703.01760 (2017).
  117. 117.Massimo Quadrana, Alexandros Karatzoglou, Balázs Hidasi, and Paolo Cremonesi. 2017. Personalizing session-based recommendations with hierarchical recurrent neural networks. In Recsys. 130–137.
  118. 118.Yogesh Singh Rawat and Mohan S Kankanhalli. 2016. ConTagNet: exploiting user context for image tag recommendation. In Proceedings of the 2016 ACM on Multimedia Conference. 1102–1106.
  119. 119.S. Rendle. 2010. Factorization Machines. In 2010 IEEE International Conference on Data Mining.
  120. 120.Steffen Rendle, Christoph Freudenthaler, Zeno Gantner, and Lars Schmidt-Thieme. 2009. BPR: Bayesian personalized ranking from implicit feedback. In Proceedings of the twenty-fifth conference on uncertainty in artificial intelligence. 452–461.
  121. 121.Francesco Ricci, Lior Rokach, and Bracha Shapira. 2015. Recommender systems: introduction and challenges. In Recommender systems handbook. 1–34.
  122. 122.Salah Rifai, Pascal Vincent, Xavier Muller, Xavier Glorot, and Yoshua Bengio. 2011. Contractive auto-encoders: Explicit invariance during feature extraction. In ICML. 833–840.
  123. 123.Ruslan Salakhutdinov, Andriy Mnih, and Geoffrey Hinton. 2007. Restricted Boltzmann machines for collaborative filtering. In ICML. 791–798.
  124. 124.Adam Santoro, David Raposo, David G Barrett, Mateusz Malinowski, Razvan Pascanu, Peter Battaglia, and Tim Lillicrap. 2017. A simple neural network module for relational reasoning. In NIPS. 4967–4976.
  125. 125.Suvash Sedhain, Aditya Krishna Menon, Scott Sanner, and Lexing Xie. 2015. Autorec: Autoencoders meet collaborative filtering. In WWW. 111–112.
  126. 126.Sungyong Seo, Jing Huang, Hao Yang, and Yan Liu. 2017. Interpretable convolutional neural networks with dual local and global attention for review rating prediction. In Recsys. 297–305.
  127. 127.Sungyong Seo, Jing Huang, Hao Yang, and Yan Liu. 2017. Representation Learning of Users and Items for Review Rating Prediction Using Attention-based Convolutional Neural Network. In MLRec.
  128. 128.Joan Serrà and Alexandros Karatzoglou. 2017. Getting deep recommenders fit: Bloom embeddings for sparse binary input/output networks. In Recsys. 279–287.
  129. 129.Ying Shan, T Ryan Hoens, Jian Jiao, Haijing Wang, Dong Yu, and JC Mao. 2016. Deep Crossing: Web-scale modeling without manually crafted combinatorial features. In SIGKDD. 255–262.
  130. 130.Xiaoxuan Shen, Baolin Yi, Zhaoli Zhang, Jiangbo Shu, and Hai Liu. 2016. Automatic Recommendation Technology for Learning Resources with Convolutional Neural Network. In International Symposium on Educational Technology. 30–34.
  131. 131.Yue Shi, Martha Larson, and Alan Hanjalic. 2014. Collaborative filtering beyond the user-item matrix: A survey of the state of the art and future challenges. ACM Computing Surveys (CSUR) 47, 1 (2014), 3.
  132. 132.Elena Smirnova and Flavian Vasile. 2017. Contextual Sequence Modeling for Recommendation with Recurrent Neural Networks. (2017).
  133. 133.Harold Soh, Scott Sanner, Madeleine White, and Greg Jamieson. 2017. Deep Sequential Recommendation for Personalized Adaptive User Interfaces. In Proceedings of the 22nd International Conference on Intelligent User Interfaces. 589–593.
  134. 134.Bo Song, Xin Yang, Yi Cao, and Congfu Xu. 2018. Neural Collaborative Ranking. arXiv preprint arXiv:1808.04957 (2018).
  135. 135.Yang Song, Ali Mamdouh Elkahky, and Xiaodong He. 2016. Multi-rate deep learning for temporal recommendation. In SIGIR. 909–912.
  136. 136.Florian Strub, Romaric Gaudel, and Jérémie Mary. 2016. Hybrid Recommender System based on Autoencoders. In Proceedings of the 1st Workshop on Deep Learning for Recommender Systems. 11–16.
  137. 137.Florian Strub and Jeremie Mary. 2015. Collaborative Filtering with Stacked Denoising AutoEncoders and Sparse Inputs. In NIPS Workshop.
  138. 138.Xiaoyuan Su and Taghi M Khoshgoftaar. 2009. A survey of collaborative filtering techniques. Advances in artificial intelligence 2009 (2009), 4.
  139. 139.Alessandro Suglia, Claudio Greco, Cataldo Musto, Marco de Gemmis, Pasquale Lops, and Giovanni Semeraro. 2017. A Deep Architecture for Content-based Recommendations Exploiting Recurrent Neural Networks. In Proceedings of the 25th Conference on User Modeling, Adaptation and Personalization. 202–211.
  140. 140.Yosuke Suzuki and Tomonobu Ozaki. 2017. Stacked Denoising Autoencoder-Based Deep Collaborative Filtering Using the Change of Similarity. In WAINA. 498–502.
  141. 141.Jiwei Tan, Xiaojun Wan, and Jianguo Xiao. 2016. A Neural Network Approach to Quote Recommendation in Writings. In Proceedings of the 25th ACM International on Conference on Information and Knowledge Management. 65–74.
  142. 142.Yong Kiam Tan, Xinxing Xu, and Yong Liu. 2016. Improved recurrent neural networks for session-based recommendations. In Recsys. 17–22.
  143. 143.Jiaxi Tang and Ke Wang. 2018. Personalized top-n sequential recommendation via convolutional sequence embedding. In WSDM. 565–573.
  144. 144.Jiaxi Tang and Ke Wang. 2018. Ranking Distillation: Learning Compact Ranking Models With High Performance for Recommender System. In SIGKDD.
  145. 145.Yi Tay, Luu Anh Tuan, and Siu Cheung Hui. 2018. Latent Relational Metric Learning via Memory-based Attention for Collaborative Ranking. In WWW.
  146. 146.Yi Tay, Anh Tuan Luu, and Siu Cheung Hui. 2018. Multi-Pointer Co-Attention Networks for Recommendation. In SIGKDD.
  147. 147.Trieu H Trinh, Andrew M Dai, Thang Luong, and Quoc V Le. 2018. Learning longer-term dependencies in rnns with auxiliary losses. arXiv preprint arXiv:1803.00144 (2018).
  148. 148.Trinh Xuan Tuan and Tu Minh Phuong. 2017. 3D Convolutional Networks for Session-based Recommendation with Content Features. In Recsys. 138–146.
  149. 149.Bartlomiej Twardowski. 2016. Modelling Contextual Information in Session-Aware Recommender Systems with Neural Networks. In Recsys.
  150. 150.Moshe Unger. 2015. Latent Context-Aware Recommender Systems. In Recsys. 383–386.
  151. 151.Moshe Unger, Ariel Bar, Bracha Shapira, and Lior Rokach. 2016. Towards latent context-aware recommendation systems. Knowledge-Based Systems 104 (2016), 165–178.
  152. 152.Benigno Uria, Marc-Alexandre Côté, Karol Gregor, Iain Murray, and Hugo Larochelle. 2016. Neural autoregressive distribution estimation. Journal of Machine Learning Research 17, 205 (2016), 1–37.
  153. 153.Aaron Van den Oord, Sander Dieleman, and Benjamin Schrauwen. 2013. Deep content-based music recommendation. In NIPS. 2643–2651.
  154. 154.Manasi Vartak, Arvind Thiagarajan, Conrado Miranda, Jeshua Bratman, and Hugo Larochelle. 2017. A Meta-Learning Perspective on Cold-Start Recommendations for Items. In Advances in Neural Information Processing Systems. 6904–6914.
  155. 155.Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Lukasz Kaiser, and Illia Polosukhin. 2017. Attention is all you need. In Advances in Neural Information Processing Systems. 5998–6008.
  156. 156.Maksims Volkovs, Guangwei Yu, and Tomi Poutanen. 2017. DropoutNet: Addressing Cold Start in Recommender Systems. In Advances in Neural Information Processing Systems. 4957–4966.
  157. 157.Jeroen B. P. Vuurens, Martha Larson, and Arjen P. de Vries. 2016. Exploring Deep Space: Learning Personalized Ranking in a Semantic Space. In Recsys.
  158. 158.Hao Wang, Xingjian Shi, and Dit-Yan Yeung. 2015. Relational Stacked Denoising Autoencoder for Tag Recommendation.. In AAAI. 3052–3058.
  159. 159.Hao Wang, Naiyan Wang, and Dit-Yan Yeung. 2015. Collaborative deep learning for recommender systems. In SIGKDD. 1235–1244.
  160. 160.Hao Wang, SHI Xingjian, and Dit-Yan Yeung. 2016. Collaborative recurrent autoencoder: Recommend while learning to fill in the blanks. In NIPS. 415–423.
  161. 161.Hao Wang and Dit-Yan Yeung. 2016. Towards Bayesian deep learning: A framework and some existing methods. TKDE 28, 12 (2016), 3395–3408.
  162. 162.Jun Wang, Lantao Yu, Weinan Zhang, Yu Gong, Yinghui Xu, Benyou Wang, Peng Zhang, and Dell Zhang. 2017. IRGAN: A Minimax Game for Unifying Generative and Discriminative Information Retrieval Models. (2017).
  163. 163.Lu Wang, Wei Zhang, Xiaofeng He, and Hongyuan Zha. 2018. Supervised Reinforcement Learning with Recurrent Neural Network for Dynamic Treatment Recommendation. In SIGKDD. 2447–2456.
  164. 164.Qinyong Wang, Hongzhi Yin, Zhiting Hu, Defu Lian, Hao Wang, and Zi Huang. 2018. Neural Memory Streaming Recommender Networks with Adversarial Training. In SIGKDD.
  165. 165.Suhang Wang, Yilin Wang, Jiliang Tang, Kai Shu, Suhas Ranganath, and Huan Liu. 2017. What Your Images Reveal: Exploiting Visual Contents for Point-of-Interest Recommendation. In WWW.
  166. 166.Xiang Wang, Xiangnan He, Liqiang Nie, and Tat-Seng Chua. 2017. Item Silk Road: Recommending Items from Information Domains to Social Users. (2017).
  167. 167.Xinxi Wang and Ye Wang. 2014. Improving content-based and hybrid music recommendation using deep learning. In MM. 627–636.
  168. 168.Xinxi Wang, Yi Wang, David Hsu, and Ye Wang. 2014. Exploration in interactive personalized music recommendation: a reinforcement learning approach. TOMM 11, 1 (2014), 7.
  169. 169.Xuejian Wang, Lantao Yu, Kan Ren, Guangyu Tao, Weinan Zhang, Yong Yu, and Jun Wang. 2017. Dynamic Attention Deep Model for Article Recommendation by Learning Human Editors’ Demonstration. In SIGKDD.
  170. 170.Jian Wei, Jianhua He, Kai Chen, Yi Zhou, and Zuoyin Tang. 2016. Collaborative filtering and deep learning based hybrid recommendation for cold start problem. IEEE, 874–877.
  171. 171.Jian Wei, Jianhua He, Kai Chen, Yi Zhou, and Zuoyin Tang. 2017. Collaborative filtering and deep learning based recommendation system for cold start items. Expert Systems with Applications 69 (2017), 29–39.
  172. 172.Jiqing Wen, Xiaopeng Li, James She, Soochang Park, and Ming Cheung. 2016. Visual background recommendation for dance performances using dancer-shared images. 521–527.
  173. 173.Caihua Wu, Junwei Wang, Juntao Liu, and Wenyu Liu. 2016. Recurrent neural network based recommendation for time heterogeneous feedback. Knowledge-Based Systems 109 (2016), 90–103.
  174. 174.Chao-Yuan Wu, Amr Ahmed, Alex Beutel, and Alexander J Smola. 2016. Joint Training of Ratings and Reviews with Recurrent Recommender Networks. (2016).
  175. 175.Chao-Yuan Wu, Amr Ahmed, Alex Beutel, Alexander J Smola, and How Jing. 2017. Recurrent recommender networks. In WSDM. 495–503.
  176. 176.Sai Wu, Weichao Ren, Chengchao Yu, Gang Chen, Dongxiang Zhang, and Jingbo Zhu. 2016. Personal recommendation using deep recurrent neural networks in NetEase. In ICDE. 1218–1229.
  177. 177.Yao Wu, Christopher DuBois, Alice X Zheng, and Martin Ester. 2016. Collaborative denoising auto-encoders for top-n recommender systems. In WSDM. 153–162.
  178. 178.Jun Xiao, Hao Ye, Xiangnan He, Hanwang Zhang, Fei Wu, and Tat-Seng Chua. 2017. Attentional factorization machines: Learning the weight of feature interactions via attention networks. arXiv preprint arXiv:1708.04617 (2017).
  179. 179.Ruobing Xie, Zhiyuan Liu, Rui Yan, and Maosong Sun. 2016. Neural Emoji Recommendation in Dialogue Systems. arXiv preprint arXiv:1612.04609 (2016).
  180. 180.Weizhu Xie, Yuanxin Ouyang, Jingshuai Ouyang, Wenge Rong, and Zhang Xiong. 2016. User Occupation Aware Conditional Restricted Boltzmann Machine Based Recommendation. 454–461.
  181. 181.Caiming Xiong, Victor Zhong, and Richard Socher. 2016. Dynamic coattention networks for question answering. arXiv preprint arXiv:1611.01604 (2016).
  182. 182.Zhenghua Xu, Cheng Chen, Thomas Lukasiewicz, Yishu Miao, and Xiangwu Meng. 2016. Tag-aware personalized recommendation using a deep-semantic similarity model with negative sampling. In CIKM. 1921–1924.
  183. 183.Zhenghua Xu, Thomas Lukasiewicz, Cheng Chen, Yishu Miao, and Xiangwu Meng. 2017. Tag-aware personalized recommendation using a hybrid deep model. (2017).
  184. 184.Hong-Jian Xue, Xinyu Dai, Jianbing Zhang, Shujian Huang, and Jiajun Chen. 2017. Deep Matrix Factorization Models for Recommender Systems.. In IJCAI. 3203–3209.
  185. 185.Carl Yang, Lanxiao Bai, Chao Zhang, Quan Yuan, and Jiawei Han. 2017. Bridging Collaborative Filtering and Semi-Supervised Learning: A Neural Approach for POI Recommendation. In SIGKDD.
  186. 186.Lina Yao, Quan Z Sheng, Anne HH Ngu, and Xue Li. 2016. Things of interest recommendation by leveraging heterogeneous relations in the internet of things. ACM Transactions on Internet Technology (TOIT) 16, 2 (2016), 9.
  187. 187.Baolin Yi, Xiaoxuan Shen, Zhaoli Zhang, Jiangbo Shu, and Hai Liu. 2016. Expanded autoencoder recommendation framework and its application in movie recommendation. In SKIMA. 298–303.
  188. 188.Haochao Ying, Liang Chen, Yuwen Xiong, and Jian Wu. 2016. Collaborative deep ranking: a hybrid pair-wise recommendation algorithm with implicit feedback. In PAKDD. 555–567.
  189. 189.Haochao Ying, Fuzhen Zhuang, Fuzheng Zhang, Yanchi Liu, Guandong Xu, Xing Xie, Hui Xiong, and Jian Wu. 2018. Sequential Recommender System based on Hierarchical Attention Networks. In IJCAI.
  190. 190.Rex Ying, Ruining He, Kaifeng Chen, Pong Eksombatchai, William L Hamilton, and Jure Leskovec. 2018. Graph Convolutional Neural Networks for Web-Scale Recommender Systems. arXiv preprint arXiv:1806.01973 (2018).
  191. 191.Wenhui Yu, Huidi Zhang, Xiangnan He, Xu Chen, Li Xiong, and Zheng Qin. 2018. Aesthetic-based clothing recommendation. In WWW. 649–658.
  192. 192.Fuzheng Zhang, Nicholas Jing Yuan, Defu Lian, Xing Xie, and Wei-Ying Ma. 2016. Collaborative knowledge base embedding for recommender systems. In SIGKDD. 353–362.
  193. 193.Qi Zhang, Jiawen Wang, Haoran Huang, Xuanjing Huang, and Yeyun Gong. Hashtag Recommendation for Multimodal Microblog Using Co-Attention Network. In IJCAI.
  194. 194.Shuai Zhang, Yi Tay, Lina Yao, and Aixin Sun. 2018. Next Item Recommendation with Self-Attention. arXiv preprint arXiv:1808.06414 (2018).
  195. 195.Shuai Zhang, Lina Yao, Aixin Sun, Sen Wang, Guodong Long, and Manqing Dong. 2018. NeuRec: On Nonlinear Transformation for Personalized Ranking. arXiv preprint arXiv:1805.03002 (2018).
  196. 196.Shuai Zhang, Lina Yao, and Xiwei Xu. 2017. AutoSVD++: An Efficient Hybrid Collaborative Filtering Model via Contractive Auto-encoders. (2017).
  197. 197.Yongfeng Zhang, Qingyao Ai, Xu Chen, and W Bruce Croft. 2017. Joint representation learning for top-n recommendation with heterogeneous information sources. In CIKM. 1449–1458.
  198. 198.Xiangyu Zhao, Long Xia, Liang Zhang, Zhuoye Ding, Dawei Yin, and Jiliang Tang. 2018. Deep Reinforcement Learning for Page-wise Recommendations. arXiv preprint arXiv:1805.02343 (2018).
  199. 199.Xiangyu Zhao, Liang Zhang, Zhuoye Ding, Long Xia, Jiliang Tang, and Dawei Yin. 2018. Recommendations with Negative Feedback via Pairwise Deep Reinforcement Learning. arXiv preprint arXiv:1802.06501 (2018).
  200. 200.Guanjie Zheng, Fuzheng Zhang, Zihan Zheng, Yang Xiang, Nicholas Jing Yuan, Xing Xie, and Zhenhui Li. 2018. DRN: A Deep Reinforcement Learning Framework for News Recommendation. In WWW. 167–176.
  201. 201.Lei Zheng, Chun-Ta Lu, Lifang He, Sihong Xie, Vahid Noroozi, He Huang, and Philip S Yu. 2018. MARS: Memory Attention-Aware Recommender System. arXiv preprint arXiv:1805.07037 (2018).
  202. 202.Lei Zheng, Vahid Noroozi, and Philip S. Yu. 2017. Joint Deep Modeling of Users and Items Using Reviews for Recommendation. In WSDM.
  203. 203.Yin Zheng, Cailiang Liu, Bangsheng Tang, and Hanning Zhou. 2016. Neural Autoregressive Collaborative Filtering for Implicit Feedback. In Recsys.
  204. 204.Yin Zheng, Bangsheng Tang, Wenkui Ding, and Hanning Zhou. 2016. A Neural Autoregressive Approach to Collaborative Filtering. In ICML.
  205. 205.Chang Zhou, Jinze Bai, Junshuai Song, Xiaofei Liu, Zhengchao Zhao, Xiusi Chen, and Jun Gao. 2017. ATRank: An Attention-Based User Behavior Modeling Framework for Recommendation. arXiv preprint arXiv:1711.06632 (2017).
  206. 206.Jiang Zhou, Cathal Gurrin, and Rami Albatal. 2016. Applying visual user interest profiles for recommendation & personalisation. (2016).
  207. 207.Fuzhen Zhuang, Dan Luo, Nicholas Jing Yuan, Xing Xie, and Qing He. 2017. Representation Learning with Pair-wise Constraints for Collaborative Ranking. In WSDM. 567–575.
  208. 208.Fuzhen Zhuang, Zhiqiang Zhang, Mingda Qian, Chuan Shi, Xing Xie, and Qing He. 2017. Representation learning via Dual-Autoencoder for recommendation. Neural Networks 90 (2017), 83–89.
  209. 209.Yi Zuo, Jiulin Zeng, Maoguo Gong, and Licheng Jiao. 2016. Tag-aware recommender systems based on deep neural networks. Neurocomputing 204 (2016), 51–60.

Citation

MLA
Zhang, S., et al. “Deep Learning Based Recommender System”. ACM Computing Surveys, vol. 52, no. 1, 2019, pp. 1–8, https://doi.org/10.1145/3285029.
APA
Zhang, S., Yao, L., Sun, A., & Tay, Y. (2019). Deep Learning Based Recommender System. ACM Computing Surveys, 52(1), 1–38. https://doi.org/10.1145/3285029
Chicago
Zhang, S., L. Yao, A. Sun, and Y. Tay. 2019. “Deep Learning Based Recommender System”. ACM Computing Surveys 52 (1): 1–38. https://doi.org/10.1145/3285029.
Harvard
Zhang, S. et al. (2019) “Deep Learning Based Recommender System”, ACM Computing Surveys, 52(1), pp. 1–38. Available at: https://doi.org/10.1145/3285029.
Vancouver
1. Zhang S, Yao L, Sun A, Tay Y (2019) Deep Learning Based Recommender System. ACM Computing Surveys 52:1–38

BibTeX

@article{Zhang_2019, title={Deep Learning Based Recommender System: A Survey and New Perspectives}, volume={52}, ISSN={1557-7341}, url={http://dx.doi.org/10.1145/3285029}, DOI={10.1145/3285029}, number={1}, journal={ACM Computing Surveys}, publisher={Association for Computing Machinery (ACM)}, author={Zhang, Shuai and Yao, Lina and Sun, Aixin and Tay, Yi}, year={2019}, month=Feb, pages={1–38} }
Metadata:Crossref

Source Code

This paper has an official code repository available. Click below to access the source code.

View Repository

Access the Paper

This paper is available from its original source. Click below to access the PDF.

Open PDF