Hybrid/Advanced Session-Based Recommender Systems

The deep learning models in SBRS which have been discussed in the previous chapters have their own strengths and weaknesses. Due to the high flexibility of deep neural networks, many neural network blocks can be integrated to construct more robust and accurate models. Many session-based recommender system utilize hybrid deep neural network models. There are also several advanced deep learning approaches that are very popular in SBRS, including graph neural networks (GNNs) and deep reinforcement learning (DRL). To this end, advanced and hybrid deep neural network models in SBRS are discussed in this chapter.