# From Fully Trained to Fully Random Embeddings: Improving Neural Machine Translation with Compact Word Embedding Tables

arXiv: https://arxiv.org/abs/2104.08677 · HTML: https://mlsanity.com/paper/2104.08677/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 18 of its 27 references that are in the map.
- First posted on arXiv 18 Apr 2021.

## Builds on (most cited first, 18 of 18)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [Rethinking the Inception Architecture for Computer Vision](https://mlsanity.com/paper/1512.00567/) (2015)
- [Distributed Representations of Words and Phrases and their Compositionality](https://mlsanity.com/paper/1310.4546/) (2013)
- [SentencePiece: A simple and language independent subword tokenizer and detokenizer for Neural Text Processing](https://mlsanity.com/paper/1808.06226/) (2018)
- [A Call for Clarity in Reporting BLEU Scores](https://mlsanity.com/paper/1804.08771/) (2018)
- [Regularizing Neural Networks by Penalizing Confident Output Distributions](https://mlsanity.com/paper/1701.06548/) (2017)
- [Using the Output Embedding to Improve Language Models](https://mlsanity.com/paper/1608.05859/) (2016)
- [OpenNMT: Open-source Toolkit for Neural Machine Translation](https://mlsanity.com/paper/1709.03815/) (2017)
- [What's Hidden in a Randomly Weighted Neural Network?](https://mlsanity.com/paper/1911.13299/) (2019)
- [Multilingual Neural Machine Translation with Knowledge Distillation](https://mlsanity.com/paper/1902.10461/) (2019)
- [Hash Embeddings for Efficient Word Representations](https://mlsanity.com/paper/1709.03933/) (2017)
- [Compressing Word Embeddings via Deep Compositional Code Learning](https://mlsanity.com/paper/1711.01068/) (2017)
- [GroupReduce: Block-Wise Low-Rank Approximation for Neural Language Model Shrinking](https://mlsanity.com/paper/1806.06950/) (2018)
- [Tensorized Embedding Layers for Efficient Model Compression](https://mlsanity.com/paper/1901.10787/) (2019)
- [Near-lossless Binarization of Word Embeddings](https://mlsanity.com/paper/1803.09065/) (2018)
- [Slim Embedding Layers for Recurrent Neural Language Models](https://mlsanity.com/paper/1711.09873/) (2017)
- [Improving Word Embedding Factorization for Compression Using Distilled Nonlinear Neural Decomposition](https://mlsanity.com/paper/1910.06720/) (2019)

## Built on by (most cited first, 2 of 2)

- [DyLoRA: Parameter Efficient Tuning of Pre-trained Models using Dynamic Search-Free Low-Rank Adaptation](https://mlsanity.com/paper/2210.07558/) (2022)
- [RAIL-KD: RAndom Intermediate Layer Mapping for Knowledge Distillation](https://mlsanity.com/paper/2109.10164/) (2021)

## Abstract (arXiv)

Embedding matrices are key components in neural natural language processing (NLP) models that are responsible to provide numerical representations of input tokens.footnoteIn this paper words and subwords are referred to as textittokens and the term textitembedding only refers to embeddings of inputs. In this paper, we analyze the impact and utility of such matrices in the context of neural machine translation (NMT). We show that detracting syntactic and semantic information from word embeddings and running NMT systems with random embeddings is not as damaging as it initially sounds. We also show how incorporating only a limited amount of task-specific knowledge from fully-trained embeddings can boost the performance NMT systems. Our findings demonstrate that in exchange for negligible deterioration in performance, any NMT model can be run with partially random embeddings. Working with such structures means a minimal memory requirement as there is no longer need to store large embedding tables, which is a significant gain in industrial and on-device settings. We evaluated our embeddings in translating English into German and French and achieved a 5.3x compression rate. Despite having a considerably smaller architecture, our models in some cases are even able to outperform state-of-the-art baselines.