Tag
This paper presents COBART, a method that fine-tunes BART with prefix control tokens to generate ad headlines with controllable length and optimized click-through rate (CTR), achieving a 25.82% improvement in ROUGE-L and 5.82% in estimated CTR over previous baselines.
LoopCTR introduces loop scaling to recommendation models, using MoE-based expert mixing and hyper-connected residuals to boost CTR prediction while allowing train-deep/infer-shallow deployment for low-latency serving.