Search papers, labs, and topics across Lattice.
3
0
3
6
UnpairGR achieves superior recommendation performance by leveraging both paired and unpaired modality observations without the need for complex feature imputation or modality-specific codebooks.
REAlign significantly reduces violations in top-ranked e-commerce products by explicitly aligning query requirements with evidence, outperforming traditional rerankers.
Relevance estimation in e-commerce can be dramatically improved by treating it as a sequential decision process, leading to a 15.94% reduction in bad-case rates.