Stacking With Auxiliary Features

TL;DR

This paper introduces a stacking method with auxiliary features that enhances ensemble performance by leveraging system provenance, achieving state-of-the-art results across diverse tasks like slot filling, entity linking, and object detection.

Contribution

It proposes a novel stacking approach that incorporates auxiliary features to better discriminate among models, improving ensemble accuracy across multiple domains.

Findings

Achieved new state-of-the-art on Cold Start Slot Filling.

Obtained new best results on Tri-lingual Entity Discovery and Linking.

Significant improvements in ImageNet object detection.

Abstract

Ensembling methods are well known for improving prediction accuracy. However, they are limited in the sense that they cannot discriminate among component models effectively. In this paper, we propose stacking with auxiliary features that learns to fuse relevant information from multiple systems to improve performance. Auxiliary features enable the stacker to rely on systems that not just agree on an output but also the provenance of the output. We demonstrate our approach on three very different and difficult problems -- the Cold Start Slot Filling, the Tri-lingual Entity Discovery and Linking and the ImageNet object detection tasks. We obtain new state-of-the-art results on the first two tasks and substantial improvements on the detection task, thus verifying the power and generality of our approach.