Generative Recommenders Introduce New Paradigm for RecSys at Scale

Generative Recommenders Introduce New Paradigm for RecSys at Scale Recommender systems are evolving from traditional embedding-similarity methods to generative objectives inspired by large language models. Recommender systems are widespread across consumer internet services but remain difficult to train and serve at scale. LLMs have driven a transition toward generative models that predict the next user action or catalog item based on historical sequences.

生成式推荐系统正在重塑大规模推荐架构 大语言模型的兴起促使推荐系统从传统的嵌入相似度匹配，转向根据用户历史序列预测下一行为或物品的生成式范式。 推荐系统虽在消费互联网中广泛应用，但在大规模训练与服务部署上难度极高。 LLM推动推荐目标发生转变，由传统的基于嵌入相似度匹配改为直接生成预测海量目录中的下一个行动或物品。

Recommender systems (RecSys) are one of the most ubiquitous machine learning problems in the consumer internet industry yet notoriously difficult to train and serve at scale. The advent of LLMs has inspired a shift from the traditional embedding-similarity-based objective to a generative one, where the goal is to predict the next action or item in a large catalog given a sequence of user histories.