Search papers, labs, and topics across Lattice.
Kuaishou Technology
6
0
7
3
Generative retrieval can achieve both shared modeling and objective-specific control, leading to significant improvements in user engagement metrics.
RGD reshapes the decoding process in generative recommendation, ensuring high-value candidates are prioritized without retraining the model.
OneRetrieval achieves the unprecedented feat of real-time editability in generative retrieval, significantly enhancing e-commerce search performance without compromising recall quality.
Reward your LLM's search queries like a discerning librarian: IG-Search uses information gain to give step-by-step feedback, boosting multi-hop reasoning without the annotation overhead.
Despite users preferring human-created videos, AI-generated content can achieve similar overall engagement on video platforms by flooding the system with sheer volume.
LLMs for e-commerce search can now be trained and evaluated on a massive, realistic dataset spanning the entire search pipeline, unlike existing datasets with anonymized data and limited scope.
