All rights reserved 1998 - 2023 ABZ ONLINE

NEWS

Neural Machine Translation (NMT) - When can neural machine translation be considered as an alternative?

This question often arises, and clients typically have their doubts. Therefore, here is a brief list of situations where you should consider discussing neural machine translation (NMT) with our project manager.

Internal Communication

Texts intended for communication within the company, such as internal emails, minutes, or intranet news, are well-suited for NMT. A light post-editing is sufficient for internal purposes since the main focus is on text comprehension.

User-generated Content in E-commerce, Support and Portals

Texts created by users or customers, such as product reviews, feedback, or comment sections under blog posts, are ideal for machine translation. They only require light post-editing to serve their purpose.

However, if newsletters, blog posts, or social media posts are part of an advertising campaign, they fall under marketing texts. In such cases, NMT should be avoided.

Product Information

Large volumes of frequently repeated text can be efficiently handled with machine translation. However, technical specifications and product information, such as ingredient lists or usage instructions, should still undergo comprehensive post-editing. This is because they provide important information for sales and decision-making.

Product information should not be confused with product descriptions, as the latter are marketing texts that focus on different aspects not considered by machines.

User Manuals

For user manuals with purely informative content, machine translation is also suitable. The prerequisite, as with all other examples, is prior deep learning training of the machine. This ensures factual correctness. Additionally, comprehensive post-editing guarantees error-free printing.

When should neural machine translation (NMT) be avoided?

Marketing Texts
 

Creative texts, especially slogans, are often rich in wordplay, idiomatic expressions, irony, and other rhetorical devices. They aim to emotionally appeal to the target audience. Neural machine translation (NMT), at its current technological state, is unable to faithfully reproduce such linguistic devices with the same meaning.

Websites and Apps
 

A website must remain intuitively usable from its homepage to the last subpage. The same applies to apps on devices. Text references must be created, and a consistent, cohesive user experience must be provided. Machines lack an overview, so caution is advised. Many website operators also differentiate between high-quality and standard-quality texts, with the latter sometimes suited for machine translation.

Regarding the adaptation of web content to a target market, the term "transcreation" is often used. It goes beyond translation and involves interpreting the text for a foreign culture's target audience, including product and service naming and color schemes. Translation engines would be completely overwhelmed in this scenario as well.

Speeches and Presentations

The transcription of spoken speeches, such as for company events or business presentations, relies on rhetorical phrasing, allusions, irony, or creative wordplay. These aspects surpass a computer's capabilities, and machine translation can yield grotesque results in such cases.

Legal Texts and Other Specialized Texts, e.g., Patents

Nested texts with long sentences, where each word is crucial, often do not lend themselves well to machine translation. Machines may lose the context and mix up individual concepts or subjects. Proofreading such translations often leads to more disadvantages than benefits, as incorrectly translated terms or even sentences can cause confusion rather than help.

Literary Texts

While it may go without saying, for the sake of completeness, we expressly advise against succumbing to the temptation of having literary texts translated by a machine.