Search papers, labs, and topics across Lattice.
This paper introduces Hint-Conditioned Generative Recommendation (HCGRec), a novel framework that enhances semantic-ID generative recommenders by addressing optimization bottlenecks during reward-based post-training. By implementing a hinting mechanism that provides minimal target-prefix hints for hard training instances, HCGRec transforms zero-reward scenarios into informative comparisons, significantly improving the model's ability to generate accurate recommendations. Experimental results demonstrate that HCGRec outperforms both supervised fine-tuning and traditional reward-based approaches, reducing the occurrence of zero-advantage training samples from over 70% to below 20%.
Transforming zero-reward training instances into valuable learning opportunities, HCGRec cuts down ineffective samples from over 70% to under 20%.
Semantic-ID generative recommenders represent each item as a short sequence of discrete semantic tokens and predict the next item by autoregressively generating this token sequence. This paradigm enables a unified generation interface for item IDs, histories, and item text, but it also creates a structured optimization bottleneck during reward-based post-training: when an early semantic token enters the wrong branch of the item-token space, finite rollout groups rarely reach the ground-truth item, so group-relative optimization receives identical zero rewards and produces no useful advantage. We propose Hint-Conditioned Generative Recommendation (HCGRec), a semantic-ID generative recommendation framework that recovers learning signal for such hard training instances. HCGRec diagnoses each instance with checkpoint rollouts and supplies a minimal target-prefix hint only when the current generator cannot reach the correct item. The model then generates the unhinted suffix under the hinted semantic branch, turning zero-reward groups into informative comparisons over item-token completions. Hinting also changes token identity: hinted prefix tokens are oracle-provided item context, while unhinted suffix tokens are sampled generation actions. We therefore introduce hint-aware credit decomposition, using supervised learning to preserve item-semantic and prefix-structure alignment for hinted tokens and GRPO to optimize the sampled suffix. Experiments on sequential recommendation benchmarks show that HCGRec substantially improves over supervised fine-tuning and vanilla reward-based post-training, while reducing zero-advantage training samples from over 70% to below 20%. The code is accessible at https://github.com/WncFht/GRec.
