# Stronger Baselines for Trustable Results in Neural Machine Translation

arXiv: https://arxiv.org/abs/1706.09733 · HTML: https://mlsanity.com/paper/1706.09733/

## Key facts

- Cited by 12 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 29 Jun 2017.

## Built on by (most cited first, 12 of 12)

- [Scientific Credibility of Machine Translation Research: A Meta-Evaluation of 769 Papers](https://mlsanity.com/paper/2106.15195/) (2021)
- [SHAPE: Shifted Absolute Position Embedding for Transformers](https://mlsanity.com/paper/2109.05644/) (2021)
- [Considerations for meaningful sign language machine translation based on glosses](https://mlsanity.com/paper/2211.15464/) (2022)
- [Attention Temperature Matters in Abstractive Summarization Distillation](https://mlsanity.com/paper/2106.03441/) (2021)
- [When Good and Reproducible Results are a Giant with Feet of Clay: The Importance of Software Quality in NLP](https://mlsanity.com/paper/2303.16166/) (2023)
- [Extended Parallel Corpus for Amharic-English Machine Translation](https://mlsanity.com/paper/2104.03543/) (2021)
- [Multilingual Neural Semantic Parsing for Low-Resourced Languages](https://mlsanity.com/paper/2106.03469/) (2021)
- [Are standard Object Segmentation models sufficient for Learning Affordance Segmentation?](https://mlsanity.com/paper/2107.02095/) (2021)
- [A Simple and Strong Baseline for End-to-End Neural RST-style Discourse Parsing](https://mlsanity.com/paper/2210.08355/) (2022)
- [Assessing the Importance of Frequency versus Compositionality for Subword-based Tokenization in NMT](https://mlsanity.com/paper/2306.01393/) (2023)
- [Regularized Adaptive Momentum Dual Averaging with an Efficient Inexact Subproblem Solver for Training Structured Neural Network](https://mlsanity.com/paper/2403.14398/) (2024)
- [Segmentation Beyond Defaults: Asymmetrical Byte Pair Encoding for Optimal Machine Translation Performance](https://mlsanity.com/paper/2511.03383/) (2025)

## Abstract (arXiv)

Interest in neural machine translation has grown rapidly as its effectiveness has been demonstrated across language and data scenarios. New research regularly introduces architectural and algorithmic improvements that lead to significant gains over "vanilla" NMT implementations. However, these new techniques are rarely evaluated in the context of previously published techniques, specifically those that are widely used in state-of-theart production and shared-task systems. As a result, it is often difficult to determine whether improvements from research will carry over to systems deployed for real-world use. In this work, we recommend three specific methods that are relatively easy to implement and result in much stronger experimental systems. Beyond reporting significantly higher BLEU scores, we conduct an in-depth analysis of where improvements originate and what inherent weaknesses of basic NMT models are being addressed. We then compare the relative gains afforded by several other techniques proposed in the literature when starting with vanilla systems versus our stronger baselines, showing that experimental conclusions may change depending on the baseline chosen. This indicates that choosing a strong baseline is crucial for reporting reliable experimental results.