# Quality expectations of machine translation

**Authors:** Andy Way  
**arXiv:** [1803.08409](https://arxiv.org/abs/1803.08409)  
**Date:** 2018-03-22  
**Summary page:** https://www.summarizepaper.com/en/paper/1803.08409v1/

## Key points

- Author Andy Way explores the deployment and evaluation of machine translation (MT) and its potential as a productivity enhancer for human translators
- Millions of people rely on MT for various use-cases on a daily basis
- There are skeptics who question the usefulness of MT
- Current deployment of MT and how its output is evaluated are discussed
- Quality in MT is no longer measured by a single 'gold standard'; context must be considered when assessing its expected lifespan or "shelf-life"
- Evaluation methods need to be enhanced and adapted to different deployment scenarios as MT technology advances
- This chapter provides insights into the utility and potential of MT while acknowledging the need for ongoing improvements in quality assessment
- Stakeholders can make informed decisions about incorporating MT into their translation workflows and maximizing its benefits.

## Summary

In the chapter titled "Quality expectations of machine translation," author Andy Way explores the deployment and evaluation of machine translation (MT) and its potential as a productivity enhancer for human translators. Millions of people rely on MT for various use-cases on a daily basis, however there are still skeptics who question its usefulness. Way addresses this issue by discussing the current deployment of MT, how its output is evaluated and ways in which its quality can be improved. He emphasizes that there is no longer a single 'gold standard' measure of quality in MT; instead, the context in which it is deployed must be taken into consideration when assessing its expected lifespan or "shelf-life". As MT technology advances, it becomes essential to enhance evaluation methods and adapt them to different deployment scenarios. This chapter provides valuable insights into the utility and potential of MT while acknowledging the need for ongoing improvements in quality assessment. By understanding these factors, stakeholders can make informed decisions about incorporating MT into their translation workflows and maximizing its benefits.