Search papers, labs, and topics across Lattice.
3
0
2
3
A single model can outperform a complex multi-stage recommendation system, achieving significant engagement boosts in live traffic.
Pretraining item embeddings can be a double-edged sword鈥攂eneficial in low-resource settings but redundant in large-scale scenarios.
Refreshing stale Semantic IDs in generative retrieval boosts recall by aligning new IDs to the existing vocabulary, sidestepping costly full model retraining.
