2 papers
cs.IR2026
Real-Time Hard Negative Sampling via LLM-based Clustering for Large-Scale Two-Tower Retrieval
Ivan Ji, Liuyi Hu, Harrison +6
The two-tower model is widely used in the retrieval stage of large-scale recommendation systems, where training typically relies on in-batch and/or out-of-batch negative sampling.…
cs.IR2026
NEXT: Reasoning-Driven Video Recommendation via a Vision-Language Model
Yuming Liu, Hongye Yang, Harrison Zhao +5
We present NEXT (Next-interest EXploration Transformer), a reasoning-driven video recommendation framework that reasons over the video a user has just watched, infers the viewer's…