A^3NCF: An Adaptive Aspect Attention Model for Rating Prediction

  title={A^3NCF: An Adaptive Aspect Attention Model for Rating Prediction},
  author={Zhiyong Cheng and Ying Ding and Xiangnan He and Lei Zhu and Xuemeng Song and M. Kankanhalli},
Current recommender systems consider the various aspects of items for making accurate recommendations. [] Key Method Specifically, we design a new topic model to extract user preferences and item characteristics from review texts. They are then used to 1) guide the representation learning of users and items, and 2) capture a user’s special attention on each aspect of the targeted item with an attention network. Through extensive experiments on several largescale datasets, we demonstrate that our model…

Figures and Tables from this paper

User Diverse Preference Modeling by Multimodal Attentive Metric Learning

A novel Multimodal Attentive Metric Learning (MAML) method, which exploits the item's multimodal features to estimate the user's special attention to different aspects of this item, and can substantially outperform the state-of-the-art baselines.

Attentive Aspect Modeling for Review-Aware Recommendation

An Attentive Aspect-based Recommendation Model (AARM) is proposed which can effectively alleviate the two aforementioned problems and significantly outperforms several state-of-the-art recommendation methods on the top-N recommendation task.

Attention-Based Adaptive Memory Network for Recommendation With Review and Rating

This paper proposes an Attention- based Adaptive Memory Network (AAMN) model to leverage historical reviews and ratings systemically and is the first to bring static features into adaptively extracting semantic information from reviews, which can not only characterize user and item from a global view, but also assist to distinguish the importance of different reviews.

Collaborative Filtering Based on Attention Mechanism

This paper proposes a collaborative filtering system based on attention mechanism and design the feature-topic model to extract the characteristics of the item from review texts and uses an attention network to get the importance of theitem’s characteristic to the user.

UGRec: Modeling Directed and Undirected Relations for Recommendation

A unified graph-based recommendation model (UGRec) is proposed, which integrates the traditional directed relations in KG and the undirected item-item co-occurrence relations simultaneously and is designed for fine-grained relation modeling.

Factor-level Attentive ICF for Recommendation

A light attention neural network is designed to integrate both item-level and factor-level attention for neural ICF models, which is model-agnostic and easy-to-implement, and shows the potential of differentiate user intents on the factor- level for ICF recommendation models.

A Capsule Network for Recommendation and Explaining What You Like and Dislike

A capsule network based model for rating prediction with user reviews, named CARP, that obtains substantial performance gain over recently proposed state-of-the-art models in terms of prediction accuracy and can successfully discover the interpretable reasons at a finer level of granularity.

Knowledge-aware recommendation model with dynamic co-attention and attribute regularize

Experimental results on two realistic datasets show that the proposed model can significantly improve recommender system effectiveness and represents an advancement beyond the compared deep models.

Feature-Level Attentive ICF for Recommendation

A general feature-level attention method for ICF models to distinguish the importance of different factors when computing the item similarity for a prediction, which is model-agnostic and easy-to-implement.

MMALFM: Explainable Recommendation by Leveraging Reviews and Images

This paper applies a proposed multi-modal aspect-aware topic model (MATM) on text reviews and item images to model users' preferences and items' features from different aspects, and estimates the aspect importance of a user towards an item.



Aspect-Aware Latent Factor Model: Rating Prediction with Ratings and Reviews

This paper applies a proposed aspect-aware topic model (ATM) on the review text to model user preferences and item features from different aspects, and estimates the aspect importance of a user towards an item, and introduces a weighted matrix to associate those latent factors with the same set of aspects discovered by ATM.

TriRank: Review-aware Explainable Recommendation by Modeling Aspects

TriRank endows the recommender system with a higher degree of explainability and transparency by modeling aspects in reviews, and allows users to interact with the system through their aspect preferences, assisting users in making informed decisions.

Ratings meet reviews, a combined approach to recommend

A unified model that combines content-based filtering with collaborative filtering, harnessing the information of both ratings and reviews is proposed, which can alleviate the cold-start problem and learn latent topics that are interpretable.

Integrating Topic and Latent Factors for Scalable Personalized Review-based Rating Prediction

An integrated topic and latent factor model (ITLFM) is proposed, which combines topics and latent factors in a linear way to make them complement each other for better accuracies in rating prediction tasks.

Attentive Group Recommendation

The AGREE model not only improves the group recommendation performance but also enhances the recommendation for users, especially for cold-start users that have no historical interactions individually.

Collaborative Multi-Level Embedding Learning from Reviews for Rating Prediction

A novel Collaborative Multi-Level Embedding model is proposed which allows CMLE to inherit the ability of capturing word local context information from word embedding model and relax the strict equivalence requirement by projecting review embedding to user and item embeddings.

TransNets: Learning to Transform for Recommendation

This work extends the DeepCoNN model by introducing an additional latent layer representing the target user-target item pair, and shows that TransNets and extensions of it improve substantially over the previous state-of-the-art performance on recommendation tasks.

Joint Representation Learning for Top-N Recommendation with Heterogeneous Information Sources

This work analyzes how information propagates among different information sources in a gradient-descent learning paradigm, and proposes an extendable version of the JRL framework (eJRL), which is rigorously extendable to new information sources to avoid model re-training in practice.

TopicMF: Simultaneously Exploiting Ratings and Reviews for Recommendation

Experimental results show the superiority of the proposed novel matrix factorization model (called TopicMF) over the state-of-the-art models, demonstrating its effectiveness for recommendation tasks.