Leveraging Low-Rank Relations Between Surrogate Tasks in Structured Prediction
This work addresses structured prediction problems for machine learning practitioners, offering an incremental improvement over previous methods by removing the need for explicit coding/decoding knowledge.
The paper tackles the problem of structured prediction by proposing an efficient algorithm based on trace norm regularization that leverages relationships between surrogate outputs without requiring explicit knowledge of coding/decoding functions, enabling application to large or infinite-dimensional surrogate spaces. Numerical experiments on ranking problems show this approach provides a significant advantage in practice.
We study the interplay between surrogate methods for structured prediction and techniques from multitask learning designed to leverage relationships between surrogate outputs. We propose an efficient algorithm based on trace norm regularization which, differently from previous methods, does not require explicit knowledge of the coding/decoding functions of the surrogate framework. As a result, our algorithm can be applied to the broad class of problems in which the surrogate space is large or even infinite dimensional. We study excess risk bounds for trace norm regularized structured prediction, implying the consistency and learning rates for our estimator. We also identify relevant regimes in which our approach can enjoy better generalization performance than previous methods. Numerical experiments on ranking problems indicate that enforcing low-rank relations among surrogate outputs may indeed provide a significant advantage in practice.