iAgent [Xu et al., ACL 2025] proposes placing an agent between a provider-side recommender system and the user. The agent searches for relevant knowledge and reranks the recommendations produced by the platform, with the goal of obtaining rankings that better reflect the user’s preferences.
The paper motivates this approach as follows:
Traditional recommender systems usually take the user-platform paradigm, where users are directly exposed under the control of the platform’s recommendation algorithms. However, the defect of recommendation algorithms may put users in very vulnerable positions under this paradigm. […] Therefore, there is an urgent need to develop a new paradigm to protect user interests and alleviate these issues.
However, this paradigm is precisely what was proposed in Private Recommender Systems: How Can Users Build Their Own Fair Recommender Systems without Log Data? [Sato, SDM 2022].
The idea was subsequently developed in works such as Towards Principled User-side Recommender Systems [Sato, CIKM 2022] and Overhead-free User-side Recommender Systems [Sato, arXiv 2024], and was later brought together in the doctoral dissertation User-Side Realization. For example, the following passage describes the paradigm:
Traditionally, recommendation algorithms have been designed for service developers. But recently, a new paradigm called user-side recommender systems has been proposed. User-side recommender systems are built and used by end users, in sharp contrast to traditional provider-side recommender systems.
Nevertheless, the paper by Xu et al. [ACL 2025] does not mention user-side recommender systems at all.
Of course, omissions of relevant prior work can happen to anyone. Still, it is particularly disheartening to see a concept that I proposed, developed through a series of papers, and eventually made the subject of my doctoral dissertation go unrecognized in a paper proposing essentially the same paradigm.
I do not know how many people this message will reach, but I can only hope that it reaches as many members of the research community as possible.
