Autoregressive Generation Strategies for Top-K Sequential Recommendations
This work addresses the challenge of predicting future user interactions in recommender systems, though it is incremental as it builds on existing transformer-based models and generation strategies.
The paper tackled the problem of improving Top-K sequential recommendations by exploring autoregressive generation strategies, including novel Reciprocal Rank Aggregation (RRA) and Relevance Aggregation (RA) methods, and found that these approaches enhance performance on longer time horizons compared to standard methods.
The goal of modern sequential recommender systems is often formulated in terms of next-item prediction. In this paper, we explore the applicability of generative transformer-based models for the Top-K sequential recommendation task, where the goal is to predict items a user is likely to interact with in the "near future". We explore commonly used autoregressive generation strategies, including greedy decoding, beam search, and temperature sampling, to evaluate their performance for the Top-K sequential recommendation task. In addition, we propose novel Reciprocal Rank Aggregation (RRA) and Relevance Aggregation (RA) generation strategies based on multi-sequence generation with temperature sampling and subsequent aggregation. Experiments on diverse datasets give valuable insights regarding commonly used strategies' applicability and show that suggested approaches improve performance on longer time horizons compared to widely-used Top-K prediction approach and single-sequence autoregressive generation strategies.