ITA: Image-Text Alignments for Multi-Modal Named Entity Recognition

TL;DR

This paper introduces ITA, a method that aligns image features into the textual space to improve multi-modal named entity recognition, enabling better interaction modeling between images and text within pretrained textual models.

Contribution

The paper proposes a novel image-text alignment approach (ITA) that enhances multi-modal NER by integrating visual information into the textual embedding space, improving interaction modeling.

Findings

Achieves state-of-the-art accuracy on multi-modal NER datasets.

Enables effective use of image information even without images.

Improves robustness to noisy or missing image data.

Abstract

Recently, Multi-modal Named Entity Recognition (MNER) has attracted a lot of attention. Most of the work utilizes image information through region-level visual representations obtained from a pretrained object detector and relies on an attention mechanism to model the interactions between image and text representations. However, it is difficult to model such interactions as image and text representations are trained separately on the data of their respective modality and are not aligned in the same space. As text representations take the most important role in MNER, in this paper, we propose {\bf I}mage-{\bf t}ext {\bf A}lignments (ITA) to align image features into the textual space, so that the attention mechanism in transformer-based pretrained textual embeddings can be better utilized. ITA first aligns the image into regional object tags, image-level captions and optical characters as visual contexts, concatenates them with the input texts as a new cross-modal input, and then feeds it into a pretrained textual embedding model. This makes it easier for the attention module of a pretrained textual embedding model to model the interaction between the two modalities since they are both represented in the textual space. ITA further aligns the output distributions predicted from the cross-modal input and textual input views so that the MNER model can be more practical in dealing with text-only inputs and robust to noises from images. In our experiments, we show that ITA models can achieve state-of-the-art accuracy on multi-modal Named Entity Recognition datasets, even without image information.