Don't Waste It: Guiding Generative Recommenders with Structured Human Priors via Multi-head Decoding
This addresses the challenge of making generative recommenders more effective and interpretable for industrial applications by leveraging existing human knowledge, though it is incremental as it builds on prior adapter-based approaches.
The paper tackles the problem of integrating structured human priors (like item taxonomies) into generative recommender systems to improve accuracy and beyond-accuracy objectives such as diversity, without discarding valuable domain knowledge. The result shows significant enhancements in experiments on three large-scale datasets, with the method also enabling better use of longer contexts and larger models.
Optimizing recommender systems for objectives beyond accuracy, such as diversity, novelty, and personalization, is crucial for long-term user satisfaction. To this end, industrial practitioners have accumulated vast amounts of structured domain knowledge, which we term human priors (e.g., item taxonomies, temporal patterns). This knowledge is typically applied through post-hoc adjustments during ranking or post-ranking. However, this approach remains decoupled from the core model learning, which is particularly undesirable as the industry shifts to end-to-end generative recommendation foundation models. On the other hand, many methods targeting these beyond-accuracy objectives often require architecture-specific modifications and discard these valuable human priors by learning user intent in a fully unsupervised manner. Instead of discarding the human priors accumulated over years of practice, we introduce a backbone-agnostic framework that seamlessly integrates these human priors directly into the end-to-end training of generative recommenders. With lightweight, prior-conditioned adapter heads inspired by efficient LLM decoding strategies, our approach guides the model to disentangle user intent along human-understandable axes (e.g., interaction types, long- vs. short-term interests). We also introduce a hierarchical composition strategy for modeling complex interactions across different prior types. Extensive experiments on three large-scale datasets demonstrate that our method significantly enhances both accuracy and beyond-accuracy objectives. We also show that human priors allow the backbone model to more effectively leverage longer context lengths and larger model sizes.