Query Driven Algorithm Selection in Early Stage Retrieval
Joel Mackenzie, J Shane Culpepper, Roi Blanco, Matt Crane, Charles LA Clarke, Jimmy Lin
WSDM'18: PROCEEDINGS OF THE ELEVENTH ACM INTERNATIONAL CONFERENCE ON WEB SEARCH AND DATA MINING | ASSOC COMPUTING MACHINERY | Published : 2018
Large scale retrieval systems often employ cascaded ranking architectures, in which an initial set of candidate documents is iteratively refined and re-ranked by increasingly sophisticated and expensive ranking models. In this paper, we propose a unified framework for predicting a range of performance-sensitive parameters based on minimizing end-to-end effectiveness loss. The framework does not require relevance judgments for training, is amenable to predicting a wide range of parameters, allows for fine tuned efficiency-effectiveness trade-offs, and can be easily deployed in large scale search systems with minimal overhead. As a proof of concept, we show that the framework can accurately pr..View full abstract
Awarded by Australian Research Council's Discovery Projects Scheme
This work was supported by the Australian Research Council's Discovery Projects Scheme (DP170102231), the Natural Sciences and Engineering Research Council of Canada, an Australian Government Research Training Program Scholarship, and a grant from the Mozilla Foundation. We thank Luke Gallagher for providing support with the Learning-to-Rank framework.