N-gram Injection into Transformers for Dynamic Language Model Adaptation in Handwritten Text Recognition

TL;DR

This paper introduces an external n-gram injection method for transformer-based handwritten text recognition models, enabling dynamic adaptation to target language distributions without retraining, thereby improving recognition accuracy on shifted corpora.

Contribution

The paper proposes a novel n-gram injection technique for real-time language model adaptation in transformers, enhancing robustness to language shifts in handwritten text recognition.

Findings

Significantly reduces performance gap on target datasets

Effective without additional training on target data

Improves robustness to language distribution shifts

Abstract

Transformer-based encoder-decoder networks have recently achieved impressive results in handwritten text recognition, partly thanks to their auto-regressive decoder which implicitly learns a language model. However, such networks suffer from a large performance drop when evaluated on a target corpus whose language distribution is shifted from the source text seen during training. To retain recognition accuracy despite this language shift, we propose an external n-gram injection (NGI) for dynamic adaptation of the network's language modeling at inference time. Our method allows switching to an n-gram language model estimated on a corpus close to the target distribution, therefore mitigating bias without any extra training on target image-text pairs. We opt for an early injection of the n-gram into the transformer decoder so that the network learns to fully leverage text-only data at the low additional cost of n-gram inference. Experiments on three handwritten datasets demonstrate that the proposed NGI significantly reduces the performance gap between source and target corpora.