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

Abstract

User reviews contain rich semantics towards the preference of users to features of items. Recently, many deep learning based solutions have been proposed by exploiting reviews for recommendation. The attention mechanism is mainly adopted in these works to identify words or aspects that are important for rating prediction. However, it is still hard to understand whether a user likes or dislikes an aspect of an item according to what viewpoint the user holds and to what extent, without examining the review details. Here, we consider a pair of a viewpoint held by a user and an aspect of an item as a logic unit. Reasoning a rating behavior by discovering the informative logic units from the reviews and resolving their corresponding sentiments could enable a better rating prediction with explanation.To this end, in this paper, we propose a capsule network based model for rating prediction with user reviews, named CARP. For each user-item pair, CARP is devised to extract the informative logic units from the reviews and infer their corresponding sentiments. The model firstly extracts the viewpoints and aspects from the user and item review documents respectively. Then we derive the representation of each logic unit based on its constituent viewpoint and aspect. A sentiment capsule architecture with a novel Routing by Bi-Agreement mechanism is proposed to identify the informative logic unit and the sentiment based representations in user-item level for rating prediction. Extensive experiments are conducted over seven real-world datasets with diverse characteristics. Our results demonstrate that the proposed CARP obtains substantial performance gain over recently proposed state-of-the-art models in terms of prediction accuracy. Further analysis shows that our model can successfully discover the interpretable reasons at a finer level of granularity.Great buiding quality, great screen, great battery life and the best feature is the Face ID, quick and realiable, much better and fast than a fingerprint scanner. The bad side its is too expensive for what it is.

Explainable Recommendation: A Survey and New Perspectives

“…The version of record is available at: http://dx.doi.org/10.1561/1500000066 Li et al (2019) developed a capsule network approach to explainable recommendation. It considers an "item aspect -user viewpoint" pair as a logic unit, which is used to reason the user rating behaviors.…”

An End-to-End Review-Based Aspect-Level Neural Model for Sequential Recommendation

“…In order to improve the accuracy and interpretability of recommendation, there is a lot of research on utilizing reviews. Both convolution neural network (CNN) [9] and recurrent neural network (RNN) [10] have been widely adopted to extract semantic representation from reviews for rating prediction [1][2][3][4][11][12][13][14]. DeepCoNN [13] is the first attempt to jointly model both the user and the item from reviews using neural networks.…”

Aspect-Based Fashion Recommendation With Attention Mechanism

“…Recently, deep learning techniques have been applied to recommender systems based on review contents with great successes. CNN [10], [11], Recurrent Neural Network (RNN) [12], [13], and Capsule Network [14] are widely used to extract the semantic contextual information [10], [15] by training the networks to learn the deep feature representation of reviews and probabilistic matrix factorization for the rating prediction. In [16] and [17], a word vector model and CNNs were used to learn users' behaviors and item's attributes.…”

“…In [18], an RNN was combined with a factorization machine via a regularization term to predict the rating of an item by using item's latent factors learned from the RNN. In [15], a sentiment capsule architecture with a novel routing called the bi-agreement mechanism was proposed to identify the informative logic unit and the sentiment-based representations in the user-item level for rating predictions.…”

“…Cheng et al designed an aspect-aware topic model via a multinomial distribution over the reviews to learn different user and item aspects in the topic space [39]. Recently, Li et al proposed an aspect-specific gating and projection to extract the viewpoint of relevant information precisely by disambiguating the semantics of each word [15].…”