Recent advancements in recommender systems are increasingly focused on enhancing efficiency and addressing biases in user interactions. Techniques like Heterogeneity-Aware Adaptive Pre-ranking are optimizing pre-ranking processes by intelligently managing sample complexity, which can lead to improved user engagement without additional computational costs. Concurrently, Time-aware Inverse Propensity Scoring is tackling selection and exposure biases in sequential recommendations, allowing for more accurate user preference modeling. The integration of generative models is also being refined, with frameworks like Personalized Semi-Autoregressive with Online Knowledge Distillation improving reranking efficiency while maintaining quality. Moreover, the emergence of frameworks such as FairLRM is bridging the gap between semantic understanding and popularity bias, enhancing recommendation fairness. As the field matures, the focus is shifting toward developing systems that not only deliver personalized content but also ensure responsible and sustainable practices, paving the way for more interactive and agentic recommender systems that adapt to real-time user dynamics.
Most large-scale recommender systems follow a multi-stage cascade of retrieval, pre-ranking, ranking, and re-ranking. A key challenge at the pre-ranking stage arises from the heterogeneity of training...
Generative models offer a promising paradigm for the final stage reranking in multi-stage recommender systems, with the ability to capture inter-item dependencies within reranked lists. However, their...
Semantic ID (SID)-based recommendation is a promising paradigm for scaling sequential recommender systems, but existing methods largely follow a semantic-centric pipeline: item embeddings are learned ...
Sequential Recommendation (SR) predicts users next interactions by modeling the temporal order of their historical behaviors. Existing approaches, including traditional sequential models and generativ...
Popularity bias is a pervasive problem in recommender systems, where recommendations disproportionately favor popular items. This not only results in "rich-get-richer" dynamics and a homogenization of...
Semantic understanding of popularity bias is a crucial yet underexplored challenge in recommender systems, where popular items are often favored at the expense of niche content. Most existing debiasin...
Model merging (MM) offers an efficient mechanism for integrating multiple specialized models without access to original training data or costly retraining. While MM has demonstrated success in domains...
Innovation in Recommender Systems is currently impeded by a fractured ecosystem, where researchers must choose between the ease of in-memory experimentation and the costly, complex rewriting required ...
Policy-based Reinforcement Learning (RL) has established itself as the dominant paradigm in generative recommendation for optimizing sequential user interactions. However, when applied to offline hist...
Modern content platforms offer paid promotion to mitigate cold start by allocating exposure via auctions. Our empirical analysis reveals a counterintuitive flaw in this paradigm: while promotion rescu...