Towards Opening the Black Box of Neural Machine Translation: Source and Target Interpretations of the Transformer

2022/05/23 by Javier Ferrando, Ferrando, Javier, Gerard I. Gállego +7 · 2 citations
Computer Science · #Computation and Language (cs.CL) #FOS: Computer and information sciences #Natural Language Processing Techniques #Text Readability and Simplification #Topic Modeling

paper · pdf · doi:10.48550/arxiv.2205.11631

openalex publication_date 2022/05/23 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

In Neural Machine Translation (NMT), each token prediction is conditioned on the source sentence and the target prefix (what has been previously translated at a decoding step). However, previous work on interpretability in NMT has mainly focused solely on source sentence tokens' attributions. Therefore, we lack a full understanding of the influences of every input token (source sentence and target prefix) in the model predictions. In this work, we propose an interpretability method that tracks input tokens' attributions for both contexts. Our method, which can be extended to any encoder-decoder Transformer-based model, allows us to better comprehend the inner workings of current NMT models. We apply the proposed method to both bilingual and multilingual Transformers and present insights into their behaviour.

Cited by

• The Unheard Alternative: Contrastive Explanations for Speech-to-Text Models
• Cross-Attention is Half Explanation in Speech-to-Text Models

Related

• On the Sub-Layer Functionalities of Transformer Decoder
• Attention Weights in Transformer NMT Fail Aligning Words Between Sequences but Largely Explain Model Predictions
• Hard but Robust, Easy but Sensitive: How Encoder and Decoder Perform in Neural Machine Translation
• Analyzing the Source and Target Contributions to Predictions in Neural\n Machine Translation
• Explaining How Transformers Use Context to Build Predictions