Pairwise Learning to Rank by Neural Networks Revisited: Reconstruction, Theoretical Analysis and Practical Performance
This work addresses ranking tasks in information retrieval, offering a simpler and more effective pairwise method, though it is incremental as it builds on existing RankNet architecture.
The authors tackled the problem of learning to rank by proposing DirectRanker, a neural network that generalizes RankNet with mathematical guarantees of reflexivity, antisymmetry, and transitivity, resulting in outperforming state-of-the-art methods on datasets like LETOR MSLR-WEB10K, MQ2007, and MQ2008.
We present a pairwise learning to rank approach based on a neural net, called DirectRanker, that generalizes the RankNet architecture. We show mathematically that our model is reflexive, antisymmetric, and transitive allowing for simplified training and improved performance. Experimental results on the LETOR MSLR-WEB10K, MQ2007 and MQ2008 datasets show that our model outperforms numerous state-of-the-art methods, while being inherently simpler in structure and using a pairwise approach only.