Sequence Adaptation via Reinforcement Learning in Recommender Systems

  title={Sequence Adaptation via Reinforcement Learning in Recommender Systems},
  author={Stefanos Antaris and Dimitrios Rafailidis},
  journal={Proceedings of the 15th ACM Conference on Recommender Systems},
Accounting for the fact that users have different sequential patterns, the main drawback of state-of-the-art recommendation strategies is that a fixed sequence length of user-item interactions is required as input to train the models. This might limit the recommendation accuracy, as in practice users follow different trends on the sequential recommendations. Hence, baseline strategies might ignore important sequential interactions or add noise to the models with redundant interactions… 

Figures and Tables from this paper

A Survey on Reinforcement Learning for Recommender Systems

A thorough overview, comparisons, and summarization of RL approaches applied in four typical recommender scenarios, including interactive recommendation, conversational recommendatin, sequential recommendation, and explainable recommendation is provided.



Self-Supervised Reinforcement Learning for Recommender Systems

This paper proposes two frameworks namely Self-supervised Q-learning and Self-Supervised Actor-Critic and integrates the proposed frameworks with four state-of-the-art recommendation models, demonstrating the effectiveness of the approach on real-world datasets.

Time Interval Aware Self-Attention for Sequential Recommendation

This paper proposes TiSASRec (Time Interval aware Self-attention based sequential recommendation), which models both the absolute positions of items as well as the time intervals between them in a sequence, which outperforms various state-of-the-art sequential models on both sparse and dense datasets and different evaluation metrics.

Session-based Recommendations with Recurrent Neural Networks

It is argued that by modeling the whole session, more accurate recommendations can be provided by an RNN-based approach for session-based recommendations, and introduced several modifications to classic RNNs such as a ranking loss function that make it more viable for this specific problem.

Sequential Recommendation with Dual Side Neighbor-based Collaborative Relation Modeling

This work proposes Sequential Collaborative Recommender (SCoRe) which effectively mines high-order collaborative information using cross-neighbor relation modeling and, additionally utilizes both user-side and item-side historical sequences to better capture user and item dynamics.

Exploring Longitudinal Effects of Session-based Recommendations

This work analyzes to what extent algorithms of different types may lead to concentration effects over time in session-based recommendation, and reveals that all investigated algorithms—both neural and heuristic ones—may lead to lower item coverage and to a higher concentration on a subset of the items.

Social Attentive Deep Q-Networks for Recommender Systems

This work develops a Social Attentive Deep Q-network (SADQN) to approximate the optimal action-value function based on the preferences of both individual users and social neighbors, by successfully utilizing a social attention layer to model the influence between them.

Contextual and Sequential User Embeddings for Large-Scale Music Recommendation

This paper proposes CoSeRNN, a neural network architecture that models users’ preferences as a sequence of embeddings, one for each session, and finds that it outperforms the current state of the art by upwards of 10% on different ranking metrics.

Adaptive Sequence Submodularity

This paper views the problem of adaptive and sequential decision making through the lens of submodularity and proposes an adaptive greedy policy with strong theoretical guarantees and runs experiments on Amazon product recommendation and Wikipedia link prediction tasks to demonstrate the practical utility of this work.

SSE-PT: Sequential Recommendation Via Personalized Transformer

A Personalized Transformer (SSE-PT) model, outperforming SASRec by almost 5% in terms of NDCG@10 on 5 real-world datasets, and can handle extremely long sequences and outperform SASRec in ranking results with comparable training speed, striking a balance between performance and speed requirements.

Where to Go Next: Modeling Long- and Short-Term User Preferences for Point-of-Interest Recommendation

This work proposes a novel method named Long- and Short-Term Preference Modeling (LSTPM) for next-POI recommendation that consists of a nonlocal network for long-term preference modeling and a geo-dilated RNN for short- term preference learning.