Can Recommender Systems Teach Themselves? A Recursive Self-Improving Framework with Fidelity Control
Luankang Zhang, Hao Wang, Zhongzhou Liu, Mingjia Yin et al.
42.08/100
💧 Filler
Padding, dubious value
Content 40.4 · Citation bonus +1.7 · 4 citations
💡 This paper proposes the Recursive Self-Improving Recommendation (RSIR) framework, which iteratively trains models on self-generated, fidelity-filtered interaction sequences to alleviate data sparsity,
#推荐系统自举#保真度过滤#隐式正则化#弱模型课程学习#数据稀疏救星#self-bootstrapping recom#fidelity filter#implicit regularizer#weak-to-strong curriculu#data sparsity mitigation
Score breakdown
Novelty5.0 / 10
Rigor3.0 / 10
Significance4.0 / 10
Clarity8.0 / 10
Reproducibility7.0 / 10
This tone hasn't been generated yet — roast it again to create it.