Abstract:While learning to rank (LTR) has been widely used in web search to prioritize most relevant webpages among the retrieved contents subject to the input queries, the traditional LTR models fail to deliver decent performance due to two main reasons: 1) the lack of well-annotated query-webpage pairs with ranking scores to cover search queries of various popularity, and 2) ill-trained models based on a limited number of training samples with poor generalization performance. To improve the performance of LTR models, tremendous efforts have been done from above two aspects, such as enlarging training sets with pseudo-labels of ranking scores by self-training, or refining the features used for LTR through feature extraction and dimension reduction. Though LTR performance has been marginally increased, we still believe these methods could be further improved in the newly-fashioned “interpolating regime”. Specifically, instead of lowering the number of features used for LTR models, our work proposes to transform original data with random Fourier feature, so as to over-parameterize the downstream LTR models (e.g., GBRank or LightGBM) with features in ultra-high dimensionality and achieve superb generalization performance. Furthermore, rather than self-training with pseudo-labels produced by the same LTR model in a “self-tuned” fashion, the proposed method incorporates the diversity of prediction results between the listwise and pointwise LTR models while co-training both models with a cyclic labeling-prediction pipeline in a “ping-pong” manner. We deploy the proposed Co-trained and Over-parameterized LTR system COLTR at Baidu search and evaluate COLTR with a large number of baseline methods. The results show that COLTR could achieve $Delta NDCG_{4}$=3.64%$sim$4.92%, compared to baselines, under various ratios of labeled samples. We also conduct a 7-day A/B Test using the realistic web traffics of Baidu Search, where we can still observe significant performance improvement around $Delta NDCG_{4}$=0.17%$sim$0.92% in real-world applications. COLTR performs consistently both in online and offline experiments.

TransRank: A Novel Algorithm for Transfer of Rank Learning

Knowledge transfer for cross domain learning to rank

A Novel Transfer Learning Method for Robot Bearing Fault Diagnosis Based on Deep Convolutional Residual Wasserstein Adversarial Network.

Transfer Learning and Deep Domain Adaptation

Scalable Heterogeneous Transfer Ranking.

CroRank: Cross Domain Personalized Transfer Ranking for Collaborative Filtering

TLRec:Transfer Learning for Cross-Domain Recommendation

Leveraging Estimated Transferability Over Human Intuition for Model Selection in Text Ranking

Pairwise cross-domain factor model for heterogeneous transfer ranking.

Cross-Stage Transfer in Multi-Stage Cascade Ranking and Filtering Systems

Transfer Learning by Ranking for Weakly Supervised Object Annotation

Query-dependent cross-domain ranking in heterogeneous network

Transfer learning based on graph ranking

Low-Rank Transfer Learning for Multi-stream Data Classification

COLTR : Semi-supervised Learning to Rank with Co-training and Over-parameterization for Web Search

Mitigating Negative Transfer in Cross-Domain Recommendation Via Knowledge Transferability Enhancement

Low-Rank Optimal Transport for Robust Domain Adaptation

Choosing Transfer Languages for Cross-Lingual Learning

A Deep Transfer Regression Method Based on Seed Replacement Considering Balanced Domain Adaptation

Cloud-based Learning System for Answer Ranking

A Simple yet Effective Framework for Active Learning to Rank