paper-with-me

홈 › Papers

Learning Language-Specific Layers for Multilingual Machine Translation

2023-05-04 · Telmo Pessoa Pires, Robin M. Schmidt, Yi-Hsiu Liao, Stephan Peitz

Multilingual Machine Translation promises to improve translation quality between non-English languages. This is advantageous for several reasons, namely lower latency (no need to translate twice), and reduced error cascades (e.g., avoiding losing gender and formality information when translating through English). On the downside, adding more languages reduces model capacity per language, which is usually countered by increasing the overall model size, making training harder and inference slower. In this work, we introduce Language-Specific Transformer Layers (LSLs), which allow us to increase model capacity, while keeping the amount of computation and the number of parameters used in the forward pass constant. The key idea is to have some layers of the encoder be source or target language-specific, while keeping the remaining layers shared. We study the best way to place these layers using a neural architecture search inspired approach, and achieve an improvement of 1.3 chrF (1.5 spBLEU) points over not using LSLs on a separate decoder architecture, and 1.9 chrF (2.2 spBLEU) on a shared decoder one.

📄 PDF Abstract BibTeX arXiv:2305.02665

Code (0)

등록된 구현이 없습니다.

Tasks

DecoderMachine TranslationNeural Architecture SearchTranslation

Methods 이 논문이 사용한 방법론

Attention 설명 없음
Adam 설명 없음
Layer Normalization Unlike batch normalization, Layer Normalization directly estimates the normalization statistics from the summed inputs…
Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.
Label Smoothing Label Smoothing is a regularization technique that introduces noise for the labels. This accounts for the fact that datasets may have mistakes in them, so maximizing the…
Dropout Dropout is a regularization technique for neural networks that drops a unit (along with connections) at training time with a specified probability $p$ (a common value is…
BPE Byte Pair Encoding, or BPE, is a subword segmentation algorithm that encodes rare and unknown words as sequences of subword units. The intuition is that various word…
Multi-Head Attention 설명 없음

Similar Papers 제목 키워드 기반

A Study of Multilingual Neural Machine Translation

2019-12-25 · Xu Tan, Yichong Leng, Jiale Chen, Yi Ren 외

Multilingual neural machine translation (NMT) has recently been investigated from different aspects (e.g., pivot translation, zero-shot translation, fine-tuning, or training from scratch) and in different settings (e.g.,…

DecoderMachine TranslationNMTTranslation

LLaVA-NeuMT: Selective Layer-Neuron Modulation for Efficient Multilingual Multimodal Translation

2025-07-25 · Jingxuan Wei, Caijun Jia, Qi Chen, Yujun Cai 외 arxiv

Multimodal Machine Translation (MMT) enhances translation quality by incorporating visual context, helping to resolve textual ambiguities. While existing MMT methods perform well in bilingual settings, extending them to …

Multimodal Machine Translation

Share or Not? Learning to Schedule Language-Specific Capacity for Multilingual Translation

2021-01-01 · ICLR 2021 1 · Biao Zhang, Ankur Bapna, Rico Sennrich, Orhan Firat

Using a mix of shared and language-specific (LS) parameters has shown promise in multilingual neural machine translation (MNMT), but the question of when and where LS capacity matters most is still under-studied. We offe…

DecoderMachine TranslationTranslation

Monolingual Adapters for Zero-Shot Neural Machine Translation

2020-11-01 · EMNLP 2020 11 · Jerin Philip, Alexandre Berard, Matthias Gall{\'e}, Laurent Besacier

We propose a novel adapter layer formalism for adapting multilingual models. They are more parameter-efficient than existing adapter layers while obtaining as good or better performance. The layers are specific to one la…

Machine TranslationTranslation

Neuron Specialization: Leveraging intrinsic task modularity for multilingual machine translation

2024-04-17 · Shaomu Tan, Di wu, Christof Monz

Training a unified multilingual model promotes knowledge transfer but inevitably introduces negative interference. Language-specific modeling methods show promise in reducing interference. However, they often rely on heu…

Cross-Lingual TransferMachine TranslationTransfer LearningTranslation