# A Reinforcement Learning Approach to Interactive-Predictive Neural Machine Translation

arXiv: https://arxiv.org/abs/1805.01553 · HTML: https://mlsanity.com/paper/1805.01553/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 3 May 2018.

## Built on by (most cited first, 8 of 8)

- [AlpacaFarm: A Simulation Framework for Methods that Learn from Human Feedback](https://mlsanity.com/paper/2305.14387/) (2023)
- [Harnessing the Power of LLMs: Evaluating Human-AI Text Co-Creation through the Lens of News Headline Generation](https://mlsanity.com/paper/2310.10706/) (2023)
- [Offline Reinforcement Learning from Human Feedback in Real-World Sequence-to-Sequence Tasks](https://mlsanity.com/paper/2011.02511/) (2020)
- [Reinforcement Learning for Generative AI: A Survey](https://mlsanity.com/paper/2308.14328/) (2023)
- [Online Learning Meets Machine Translation Evaluation: Finding the Best Systems with the Least Human Effort](https://mlsanity.com/paper/2105.13385/) (2021)
- [Interactive Machine Learning: A State of the Art Review](https://mlsanity.com/paper/2207.06196/) (2022)
- [Onception: Active Learning with Expert Advice for Real World Machine Translation](https://mlsanity.com/paper/2203.04507/) (2022)
- [One Framework to Rule Them All: Unifying RL-Based and RL-Free Methods in RLHF](https://mlsanity.com/paper/2503.19523/) (2025)

## Abstract (arXiv)

We present an approach to interactive-predictive neural machine translation that attempts to reduce human effort from three directions: Firstly, instead of requiring humans to select, correct, or delete segments, we employ the idea of learning from human reinforcements in form of judgments on the quality of partial translations. Secondly, human effort is further reduced by using the entropy of word predictions as uncertainty criterion to trigger feedback requests. Lastly, online updates of the model parameters after every interaction allow the model to adapt quickly. We show in simulation experiments that reward signals on partial translations significantly improve character F-score and BLEU compared to feedback on full translations only, while human effort can be reduced to an average number of 5 feedback requests for every input.