Back to results
US 2025/0061117 A1Published application
Cross-List Learning to Rank
AssigneeGoogle Inc.
InventorGil Shamir
Life of this patent
Estimating the term…
Abstract
Provided are systems and methods that perform learning to rank using training data for two or more different training lists. Specifically, a training dataset can include a number of training examples. Each training example can include a query and a plurality of items that are potentially responsive to the query. The ranking model can be trained using pairs of items taken from two different training examples.
Claims20
A computer-implemented method to perform cross-list learning to rank, the method comprising:
- obtaining, by a computing system comprising one or more computing devices, a first training example and a second, different training example, the first training example comprising a first plurality of items and a first query, and the second training example comprising a second plurality of items and a second query;
- processing, by the computing system, a first item from the first plurality of items with a ranking model to generate a first intermediate representation for the first item;
- processing, by the computing system, a second item from the second plurality of items with the ranking model to generate a second intermediate representation for the second item;
- determining, by the computing system, a correlation score between the first query and the second query;
- evaluating, by the computing system, a weighted pairwise ranking loss based on the first intermediate representation, the second intermediate representation, the correlation score, a first label associated with the first item, and a second label associated with the second item;
- and
- modifying, by the computing system, the ranking model based on the weighted pairwise ranking loss.
4 more elements
One or more non-transitory computer-readable media that store computer-readable instructions that, when executed by a computing system, cause the computing system to perform operations, the operations comprising:
- obtaining, by the computing system, a first training example and a second, different training example, the first training example comprising a first plurality of items and a first query, and the second training example comprising a second plurality of items and a second query;
- processing, by the computing system, a first item from the first plurality of items with a ranking model to generate a first intermediate representation for the first item;
- processing, by the computing system, a second item from the second plurality of items with the ranking model to generate a second intermediate representation for the second item;
- determining, by the computing system, a correlation score between the first training example and the second training example based at least in part on a first attribute associated with the first training example and a second attribute associated with the second training example;
- evaluating, by the computing system, a weighted pairwise ranking loss based on the first intermediate representation, the second intermediate representation, the correlation score, a first label associated with the first item, and a second label associated with the second item;
- and
- modifying, by the computing system, the ranking model based on the weighted pairwise ranking loss.
4 more elements
A computing system to perform cross-list learning to rank to distill from a teacher ranking model to a student ranking model, the computing system configured to perform operations, the operations comprising:
- obtaining, by the computing system, a first training example and a second, different training example, the first training example comprising a first plurality of items and a first query, and the second training example comprising a second plurality of items and a second query;
- processing, by the computing system, a first item from the first plurality of items with the teacher ranking model to generate a first teacher intermediate representation for the first item;
- processing, by the computing system, a second item from the second plurality of items with the teacher ranking model to generate a second teacher intermediate representation for the second item;
- processing, by the computing system, the first item from the first plurality of items with the student ranking model to generate a first student intermediate representation for the first item;
- processing, by the computing system, the second item from the second plurality of items with the student ranking model to generate a second student intermediate representation for the second item;
- determining, by the computing system, a correlation score between the first query and the second query;
- evaluating, by the computing system, a weighted pairwise ranking loss based on the first teacher intermediate representation, the second teacher intermediate representation, the first student intermediate representation, the second student intermediate representation, the correlation score, a first label associated with the first item, and a second label associated with the second item;
- and
- modifying, by the computing system, at least the student ranking model based on the weighted pairwise ranking loss.
6 more elements
Dependencies are inferred from the claim text.