paper-with-me

Papers

AdMix: A Mixed Sample Data Augmentation Method for Neural Machine Translation

2022-05-10 · Chang Jin, Shigui Qiu, Nini Xiao, Hao Jia

In Neural Machine Translation (NMT), data augmentation methods such as back-translation have proven their effectiveness in improving translation performance. In this paper, we propose a novel data augmentation approach for NMT, which is independent of any additional training data. Our approach, AdMix, consists of two parts: 1) introduce faint discrete noise (word replacement, word dropping, word swapping) into the original sentence pairs to form augmented samples; 2) generate new synthetic training data by softly mixing the augmented samples with their original samples in training corpus. Experiments on three translation datasets of different scales show that AdMix achieves signifi cant improvements (1.0 to 2.7 BLEU points) over strong Transformer baseline. When combined with other data augmentation techniques (e.g., back-translation), our approach can obtain further improvements.

📄 PDF Abstract BibTeX arXiv:2205.04686

Code (0)

등록된 구현이 없습니다.

Tasks

Data AugmentationMachine TranslationNMTSentenceTranslation

Methods 이 논문이 사용한 방법론

Multi-Head Attention 설명 없음
Attention 설명 없음
Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.
Layer Normalization Unlike batch normalization, Layer Normalization directly estimates the normalization statistics from the summed inputs…
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…
Absolute Position Encodings Absolute Position Encodings are a type of position embeddings for [Transformer-based models] where positional encodings are…
Residual Connection 설명 없음
Position-Wise Feed-Forward Layer 설명 없음

Similar Papers 제목 키워드 기반

GradMix: Gradient-based Selective Mixup for Robust Data Augmentation in Class-Incremental Learning

2025-05-13 · Minsu Kim, Seong-Hyeon Hwang, Steven Euijong Whang

In the context of continual learning, acquiring new knowledge while maintaining previous knowledge presents a significant challenge. Existing methods often use experience replay techniques that store a small portion of p…

class-incremental learningClass Incremental LearningContinual LearningData Augmentation+1

A note on the distribution of admixture segment lengths and ancestry proportions under pulse and two-wave admixture models

2015-09-19

Admixed populations are formed by the merging of two or more ancestral populations, and the ancestry of each locus in an admixed genome derives from either source. Consider a simple "pulse" admixture model, where populat…

On the number of genealogical ancestors tracing to the source groups of an admixed population

2022-10-22 · Jazlyn A. Mooney, Lily Agranat-Tamir, Jonathan K. Pritchard, Noah A. Rosenberg

In genetically admixed populations, admixed individuals possess ancestry from multiple source groups. Studies of human genetic admixture frequently estimate ancestry components corresponding to fractions of individual ge…

Rethinking Mixup for Improving the Adversarial Transferability

2023-11-28 · Xiaosen Wang, Zeyuan Yin

Mixup augmentation has been widely integrated to generate adversarial examples with superior adversarial transferability when immigrating from a surrogate model to other models. However, the underlying mechanism influenc…

Human genetic admixture through the lens of population genomics

2021-09-24 · Shyamalika Gopalan, Samuel Patillo Smith, Katharine Korunes, Iman Hamid 외

Over the last fifty years, geneticists have made great strides in understanding how our species' evolutionary history gave rise to current patterns of human genetic diversity classically summarized by Lewontin in his 197…

Diversity