Cross-Modal Robustness Transfer (CMRT): Training Robust Speech Translation Models Using Adversarial Text
This addresses the problem of robustness in speech translation for real-world applications like non-native or dialectal speech, offering a novel solution to a specific bottleneck.
The paper tackles the vulnerability of end-to-end speech translation models to adversarial attacks on inflectional morphology, proposing Cross-Modal Robustness Transfer (CMRT) to transfer robustness from text to speech without adversarial speech data, resulting in an average improvement of over 3 BLEU points across four language pairs.
End-to-End Speech Translation (E2E-ST) has seen significant advancements, yet current models are primarily benchmarked on curated, "clean" datasets. This overlooks critical real-world challenges, such as morphological robustness to inflectional variations common in non-native or dialectal speech. In this work, we adapt a text-based adversarial attack targeting inflectional morphology to the speech domain and demonstrate that state-of-the-art E2E-ST models are highly vulnerable it. While adversarial training effectively mitigates such risks in text-based tasks, generating high-quality adversarial speech data remains computationally expensive and technically challenging. To address this, we propose Cross-Modal Robustness Transfer (CMRT), a framework that transfers adversarial robustness from the text modality to the speech modality. Our method eliminates the requirement for adversarial speech data during training. Extensive experiments across four language pairs demonstrate that CMRT improves adversarial robustness by an average of more than 3 BLEU points, establishing a new baseline for robust E2E-ST without the overhead of generating adversarial speech.