Linking Representations with Multimodal Contrastive Learning

TL;DR

This paper introduces CLIPPINGS, a multimodal contrastive learning approach that aligns vision and language embeddings to improve record linkage, especially in noisy OCR scenarios, outperforming traditional string matching methods.

Contribution

The study develops a novel multimodal contrastive learning framework that leverages both images and OCR texts for improved record linkage, demonstrating significant performance gains over existing methods.

Findings

CLIPPINGS outperforms string matching in linking historical Japanese firms.

Self-supervised multimodal models surpass traditional string matching techniques.

Multimodal pre-training enhances vision-only encoder performance even with single modality at inference.

Abstract

Many applications require linking individuals, firms, or locations across datasets. Most widely used methods, especially in social science, do not employ deep learning, with record linkage commonly approached using string matching techniques. Moreover, existing methods do not exploit the inherently multimodal nature of documents. In historical record linkage applications, documents are typically noisily transcribed by optical character recognition (OCR). Linkage with just OCR'ed texts may fail due to noise, whereas linkage with just image crops may also fail because vision models lack language understanding (e.g., of abbreviations or other different ways of writing firm names). To leverage multimodal learning, this study develops CLIPPINGS (Contrastively LInking Pooled Pre-trained Embeddings). CLIPPINGS aligns symmetric vision and language bi-encoders, through contrastive language-image pre-training on document images and their corresponding OCR'ed texts. It then contrastively learns a metric space where the pooled image-text embedding for a given instance is close to embeddings in the same class (e.g., the same firm or location) and distant from embeddings of a different class. Data are linked by treating linkage as a nearest neighbor retrieval problem with the multimodal embeddings. CLIPPINGS outperforms widely used string matching methods by a wide margin in linking mid-20th century Japanese firms across financial documents. A purely self-supervised model - trained only by aligning the embeddings for the image crop of a firm name and its corresponding OCR'ed text - also outperforms popular string matching methods. Fascinatingly, a multimodally pre-trained vision-only encoder outperforms a unimodally pre-trained vision-only encoder, illustrating the power of multimodal pre-training even if only one modality is available for linking at inference time.