Content recommender systems (RSs), which recommend content (e.g., news, videos) to users, can inadvertently facilitate the spread of misinformation (e.g., fake news, inaccurate content) due to their ignorance of content credibility. The widespread misinformation can lead to serious consequences, including public confusion and social unrest. Despite the pressing need to mitigate misinformation in content RSs, only a few studies have attempted to tackle this. Moreover, they generally require retraining the entire model from scratch, which is time-consuming and impractical for real-world applications. To this end, we propose MisEraser , a novel misinformation unlearning framework that effectively mitigates misinformation in content RSs without retraining. Specifically, MisEraser consists of three key components: (1) A misinformation disentanglement network , which effectively separates misinformation-specific information from item content while preserving misinformation-free information to maintain recommendation accuracy; (2) A misinformation fusion network , which fuses misinformation-specific signals from both users’ interaction sequences and the entire misinformation item set to enhance the comprehensiveness of misinformation unlearning; and (3) A misinformation erasing network , which effectively removes the fused misinformation-specific signals from trained recommendation models, enabling them to generate misinformation-suppressed and more responsible recommendations. Extensive experiments demonstrate the effectiveness of MisEraser in mitigating misinformation while maintaining high recommendation accuracy and efficiency.

Sequential recommender systems (SRSs) aim to recommend the next items to well match users' preferences. In addition to recommendation accuracy, diversity is another critical aspect in evaluating SRSs. Recently, the emerging diffusion models (DMs) have been widely adopted in SRSs. Their employed learning-to-generate paradigm allows them to cover a much broader range of users' preferences and thus generate more diversified items. However, existing DM-based SRSs still face two significant gaps that prevent them from further improving the recommendation diversity: (1) they often rely on non-diversified users' preferences as guidance to direct the training of diffusion networks, restricting networks' ability to generate diverse items; and (2) they are based on a homogeneous diffusion inference mechanism to generate the next items and thus can only accommodate users' major preferences. Such a practice neglects users' heterogeneous preferences towards various types of items, further limiting recommendation diversity. To bridge these two critical gaps and to further unleash the potential of DMs in enhancing the recommendation diversity of SRSs, we propose a novel diversity-guided diffusion model for sequential recommendations, called DiffDiv for short. To be specific, first, a new diversity-aware guidance learning mechanism is devised to direct the training of DMs to effectively capture users' diversified preferences from their historical interactions. Then, a novel heterogeneous diffusion inference mechanism is designed to generate diversified items to accommodate users' heterogeneous preferences, further boosting the recommendation diversity. Extensive experiments on real-world datasets validate the effectiveness of DiffDiv in terms of both recommendation accuracy and diversity.

Recommender systems (RS) have found extensive applications in web and internet services, yielding significant advantages for both users and service providers, including domains such as E-commerce and Movie platforms, News portals and so on. As a crucial branch of recommendation services, multibehavior recommendation aims to leverage auxiliary behavior data (e.g., page-view and add-to-cart) to improve the recommendation performance on target behavior (e.g., purchase). Multibehavior recommendation align more closely with real-world scenarios, resulting greater research significance. However, there are two-fold issues existed in the present models that adversely affect the service quality of multibehavior RS: 1) the behavior overlap (i.e., the interaction records between identical user-item pairs co-existed in different behavior data) hinders accurate user preference learning; and 2) the commonalities of user preferences over items under different behaviors are not well explored, thereby making behavior intercorrelations not been fully captured. To handle these, we propose a nonoverlapping heterogeneous graph multibehavior recommendation (No-HGMR) model. To tackle the first problem, we erase the overlap between multitype behavior data to construct a nonoverlapping heterogeneous graph. In this graph, each specific user-item pair is connected by only one type of edge, mitigating confusion and enabling more accurate user preference learning. To solve the second problem, we construct correlation features based on user properties and item attributes. These features are then incorporated into the user preference learning procedure, facilitating the understanding of users’ common preference over items under different behaviors. Extensive experiments on real-world datasets verify the effectiveness of No-HGMR as compared to competitive state-of-the-art methods.