Overview / Usage

Key to recommender systems is learning user preferences, which are expressed through various modalities. In online reviews, for instance, this manifests in numerical rating, textual content, as well as visual images. We hypothesize that modelling these modalities jointly would result in a more holistic representation of a review towards more accurate recommendations.

Methodology / Approach

We design the MRG model, which jointly models rating prediction and text generation at the review level by incorporating LSTM cells with a novel fusion gate as a kind of soft attention to weigh the relative contributions of sentiment features and visual features that provide context to the text generation. We also derive the learning and inference algorithms respectively.

Technologies Used

Hardware: Intel Xeon Processors
Software: Python, TensorFlow

Repository

https://github.com/PreferredAI/mrg

Other links

• Paper
• Homepage