Computational and Statistical Tradeoffs in Learning to Rank
arXiv:1608.06203
Abstract
For massive and heterogeneous modern datasets, it is of fundamental interest to provide guarantees on the accuracy of estimation when computational resources are limited. In the application of learning to rank, we provide a hierarchy of rank-breaking mechanisms ordered by the complexity in thus generated sketch of the data. This allows the number of data points collected to be gracefully traded off against computational resources available, while guaranteeing the desired level of accuracy. Theoretical guarantees on the proposed generalized rank-breaking implicitly provide such trade-offs, which can be explicitly characterized under certain canonical scenarios on the structure of the data.
30 pages 5 figures
References in corpus (6)
- Spectral MLE: Top- Rank Aggregation from Pairwise Comparisons
- Improved Sum-of-Squares Lower Bounds for Hidden Clique and Hidden Submatrix Problems
- Tradeoffs for Space, Time, Data and Risk in Unsupervised Learning
- Rank Centrality: Ranking from Pair-wise Comparisons
- Oracle inequalities for computationally adaptive model selection
- Sum-of-squares lower bounds for planted clique