Web-Search Ranking with Initialized Gradient Boosted Regression Trees

By Ananth Mohan, Zheng Chen, Kilian Q. Weinberger. 2010

165 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

1. Yoshua Bengio
2. Han-Jia Ye
3. Quoc V. Le
4. Chuan Guo
5. Yu Sun
6. Gao Huang
7. Fei Sha
8. Trevor Darrell

Citations in the corpus

• Making Tree Ensembles Interpretable (2016, background): “Indeed, they have attained high scores in many data mining competitions such as web ranking (Mohan et al., 2011).”
• Supervised Word Mover's Distance (2016, background): “Document distances are a key component of many text retrieval tasks such as web-search ranking 21 [24], book recommendation [16], and news categorization [25].”
• Optimal Action Extraction for Random Forests and Boosted Trees (2015, methodology): “Often referred to as the best off-the-shelf classifiers [16], ATMs have been widely deployed into many industrial products such as Kinect [26] and face detection in camera [29], and are the must-try methods for lots of data mining competitions such as web search ranking [25].”
• Overview of random forest methodology and practical guidance with emphasis on computational biology and bioinformatics (2012, background): “…[75] yes yes all ∗ F90 Random forest [9] yes no all ∗ F77 Randomforest-matlab [39] yes no all ∗ C/C++/ randomForest-R package [45] no yes all C++/S Random Jungle [64] yes yes † Win/Unix C++ RT-Rank [53] yes yes Unix ∗ C++/Python Waffles [28] no no Win/Unix/Mac C++ Weka 3 [33] no no all 3 and 4.”
• A cross-benchmark comparison of 87 learning to rank methods (2015, unknown): “The study reported evaluation results in a different metric than the metrics chosen for this meta-analysis (Kersting & Xu, 2009; Mohan, Chen, & Weinberger, 2011; Pahikkala, Tsivtsivadze, Airola, Järvinen, & Boberg, 2009; Thuy, Vien, Viet, & Chung, 2009; Yu & Joachims, 2009).”
• Adding semantics to microblog posts (2012, background): “In recent years, gradient boosted regression trees (GBRTs) have been established as the de facto state-of-the-art learning paradigm for web search ranking [7, 12, 29].”
• Learning (predictive) risk scores in the presence of censoring due to interventions (2015, methodology): “We chose to extend boosted regression trees as this is one of the most widely used algorithms for ranking (e.g., see Mohan et al, 2011).”
• Quality versus efficiency in document scoring with learning-to-rank models (2016, unknown): “Mohan, Chen, and Weinberger (2011) propose to initialize GBRT with the predictions produced by a previously learned RF ranker.”
• Efficient Cost-Aware Cascade Ranking in Multi-Stage Retrieval (2017, methodology): “� is cascade model, however, cannot be used with gradient-boosted tree models, which are now widely believed to be state-of-the art for web search ranking algorithms [25, 30].”