Post Editing

Machine Translation Post Editing (MTPE)
– On-Demand

Need Post-Editing Service?

mt-post-editing-chart

The Best MTPE Solution Just Got Even Better

Stepes has greatly simplified machine translation post-editing by significantly improving the efficiency of the process. Our patent-pending mobile solution automatically collects post-editing feedback that is critical for MT training. These metrics include mistranslation, missing translation, grammar and spelling mistakes, terminology errors, unteligable text, time spent editing each segment pair, percentage of changes, and readability. This editing data can be fed back to the MT engine for machine learning and MT training to continuously improve quality. More importantly, Stepes MT post-editing solutions are the first to be available on mobile platforms, greatly simplifying the post-editing process into an easy chat interface.

MT Post Editing Simplified

MT Post Editing Simplified

Stepes revolutionizes machine translation post-editing (MTPE) so human linguistic reviewers and subject matter experts can confidently edit MT results easily on mobile devices, in addition to desktop computers. With the rapid development of neural MT engines powered by big data, machine translation post-editing services are increasingly needed to ensure linguistic quality and consistency. MTPE is now a standard step in the professional translation and localization process that delivers unrivaled efficiency and reduced costs.

An important application of post-editing is the ability to train MT algorithms for enhanced technical accuracy. In addition to relying on professional translators, MT developers sometimes need to involve subject matter experts such as bilingual medical professionals, legal experts, and technical engineers to post-edit MT results in order to train their automatic translation engines through machine learning. However, traditional translation software like CAT tools are archaic and difficult to use, unable to capture editing intelligence such as percentage of change or editing time per segment. However, with Stepes patented mobile translation solution MTPE is easy and accessible by industry experts and professional linguists alike, anywhere and anytime, while providing MT developers with sophisticated post-editing intelligence.

Given today’s large-scale MT training requirements, companies often need many translators (sometimes hundreds of translators working together) to edit millions of words quickly in order to reduce time-to-market and gain a competitive advantage. Stepes mobile post-editing solution allows potentially thousands of human translators to work simultaneously through our cloud-based translation ecosystem for unprecedented speed and scalability.

How It Works

How It Works

Stepes simplifies MTPE into an intuitive chat-based interface. A pair of sentences in the source and target language are shown in text bubbles for the linguist to review while changes are then entered into a 3rd bubble with a different color code. For reporting purposes, the human reviewer can also enter comments into a separate comment field. The entire process is similar to chatting on a mobile phone, making the process highly efficient for smartphone users.

Finally, A Scalable MT Post-Editing Solution

With mobile-based machine translation post-editing solutions, companies are finally able to achieve critical mass by executing large scale MTPE of machine translated content to meet enterprise demands on a global scale. Companies like Amazon and Facebook can finally translate all of their user generated content, as it is continuously being created, using the same crowd-sourced model that produced the original content.

For example, there are thousands of products added to Amazon’s platform in India. Stepes mobile translation solutions allows Amazon to simultaneously translate the product descriptions into all 23 major Indian languages with potentially thousands of Indian translators working from their smartphones.

Machine Translation (MT) Training

Case Study:
Machine Translation (MT) Training

One major factor affecting MT development is the availability of a large number of bilingual sentences to train computer algorithms (including the latest neural MT) to output quality results. However, the challenge of building such a large collection of quality translated sentences quickly is nearly impossible with conventional translation models. This is where Stepes’ Big Translation is needed to enable a large number of bilingual experts to simultaneously translate a vast number of sentences with speed and quality. Stepes is able to efficiently translate millions of sentences for MT training on-demand using thousands of pre-screened translators working seamlessly on their smartphones.

stepes-support-team-member-3

Need high volume machine translation post-editing (MTPE)?

Contact us today for a live demo and a customized solution!

  • Extremely fast project turnaround.
  • Unprecedented post-editing scalability.
  • Big Translation for machine translation (MT) training.
  • Higher quality with bilingual subject matter experts.
  • Competitive service through efficiency.
  • Mobile linguistic review enabled.