# Neural System Combination for Machine Translation

arXiv: https://arxiv.org/abs/1704.06393 · HTML: https://mlsanity.com/paper/1704.06393/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 21 Apr 2017.

## Built on by (most cited first, 6 of 6)

- [SpanNER: Named Entity Re-/Recognition as Span Prediction](https://mlsanity.com/paper/2106.00641/) (2021)
- [RefSum: Refactoring Neural Summarization](https://mlsanity.com/paper/2104.07210/) (2021)
- [Hit the Sweet Spot! Span-Level Ensemble for Large Language Models](https://mlsanity.com/paper/2409.18583/) (2024)
- [Stick to Facts: Towards Fidelity-oriented Product Description Generation](https://mlsanity.com/paper/2503.08454/) (2025)
- [Statistical Machine Translation for Indic Languages](https://mlsanity.com/paper/2301.00539/) (2023)
- [Multilingual KokoroChat: A Multi-LLM Ensemble Translation Method for Creating a Multilingual Counseling Dialogue Dataset](https://mlsanity.com/paper/2603.22913/) (2026)

## Abstract (arXiv)

Neural machine translation (NMT) becomes a new approach to machine translation and generates much more fluent results compared to statistical machine translation (SMT). However, SMT is usually better than NMT in translation adequacy. It is therefore a promising direction to combine the advantages of both NMT and SMT. In this paper, we propose a neural system combination framework leveraging multi-source NMT, which takes as input the outputs of NMT and SMT systems and produces the final translation. Extensive experiments on the Chinese-to-English translation task show that our model archives significant improvement by 5.3 BLEU points over the best single system output and 3.4 BLEU points over the state-of-the-art traditional system combination methods.