Post-publishing: the future of translation

Post-editing is becoming an essential part of the translation process. It is revolutionizing our approach to content translation in the digital age, combining artificial intelligence and human expertise to meet the growing need for fast, high-quality translation.

What is post-editing?

Post-editing consists of revising and improving a text previously translated by a machine translation system. A professional linguist refines the raw result generated by the artificial intelligence, thus combining technological efficiency with human expertise.

Historical perspective

The evolution of post-editing is closely linked to that of machine translation. From initial mixed experiences in the 1950s-1960s, we moved on to the emergence of computer-aided translation (CAT) systems in the 1980s. The 1990s saw the development of translation memories, followed by the appearance of the first professional post-editing services in the early 2000s.

In addition, the launch of Google Translate in 2006 democratized access to machine translation. The advent of neural machine translation in 2016 significantly improved the quality of raw translations. Today, the development ofAutomatic Post-Editing (APE) and the growing integration of AI mark a new era in this field.

The different types of post-editing

There are three main types of post-editing: light post-editing, full post-editing and Automatic Post-Editing (APE).

Light post-editing focuses on correcting serious errors,eliminating obvious mistakes and checking basic terminology.

The complete post-edition goes a step further by improving style, adapting content culturally and harmonizing terminology.

The APE represents a significant step forward in automating the correction process. It uses machine learning algorithms and corpora of post-edited human translations to continuously improve its performance.

The post-editing process

The post-editing process comprises three main stages: machine translation of the source text, revision and correction by a human post-editor, and final quality control.

Machine translation tools such as Google Translate, DeepL and Microsoft Translator provide the basis for this. Post-editing tools such as SDL Trados Studio, MemoQ and MateCat facilitate the work of post-editors. Finally, APE-specific tools such as ModernMT and Systran Pure Neural Server further automate the process.

Post-publishing: the future of translation in the digital age

Benefits and challenges

Post-editing offers many advantages, including significant time and cost savings. It enables larger volumes to be processed in a shorter timeframe, while reducing costs by 30% to 50% compared to traditional translation. In addition, terminology consistency is also improved through the use of translation memories and terminology databases.

The productivity of translators is considerably increased with post-editing. An experienced post-editor can process an average of 3500-5000 words per day, compared to 2000-3000 words for a traditional translation. Studies have shown that post-editing can increase productivity by up to 66%, or even 100% for technical content.

However, post-editing also presents challenges. The variable quality of machine translation may require in-depth revision for some texts. There is a risk of linguistic contamination, where the post-editor can be influenced by the structures of the source language. Adaptation to cultural specificities remains a challenge, as cultural nuances can escape machine translation systems.

Economic implications of post-publishing

For companies, post-editing represents an opportunity to significantly reduce translation costs and accelerate delivery times. Theinitial investmentin machine translation and post-editing tools can quickly pay for itself, especially for companies with regular translation needs.

For freelance translators, post-editing entails an evolution of the remuneration model, often moving from word-based to hourly oreffort-based post-editing. This evolution requires training and offers opportunities for specialization in specific technical fields. However, increased competition can exert downward pressure on rates.

The future of post-editing

The future of post-editing looks bright, with machine translation engines continuing to improve thanks to advances inAI. We can expect post-editors to become increasingly specialized in specific fields, adding maximum value.

The integration ofmachine learning will enable systems to continually learn from post-editors’ corrections. The development of automated quality assurance will help to detect and correct common errors more effectively.

There will probably be a shift towards pre-editing, where theoptimization of source texts for machine translation will gain in importance. Post-editing systems will increasingly adapt to specific customer and industry preferences, offering greater customization.

Ethical considerations in post-editing

The rise of post-editing raises important ethical questions. How can high standards be maintained in the face of productivity pressures? Who is responsible for the final content – the AI, the post-editor, or both? Transparency towards end-customers regarding the use of post-editing is also a matter of debate.

In addition, the impact on employment in the translation sector in the face ofincreasing automation is a major concern. These issues call for in-depth reflection to ensure the ethical and sustainable development of post-editing.

AFTraduction was a proud participant in the CNET (Chambre Nationale des Entreprises de Traduction) conference on June 14, 2014. CNET entitled the event ” AI and translation: deciphering the issues and seizing the opportunities “. The event brought together numerous experts from the translation industry. Participants discussed the ethical issues raised by the use of post-editing, and examined the challenges and opportunities it creates for the industry. As a result, post-editing will undoubtedly continue to drive debate in the years to come, underlining its growing importance in the professional translation landscape.

Conclusion

Post-editing is establishing itself as an essential approach in the translation industry, offering a balance between the efficiency of machine translation and the finesse of human expertise. Despite persistent challenges, technological advances and the adaptation of professionals promise a promising future for this discipline.

Companies and translators who have mastered the art of post-editing are better positioned to meet the growing demand for fast, high-quality translations. The key to success will lie in the ability to judiciously combine technology and human expertise, while remaining attentive to the ethical and qualitative considerations inherent in this constantly evolving field.