Training MT Engines to Deliver High-Quality Global Content, Fast

welocalize January 27, 2021
McAfee & Welocalize Case Study

The largest independent pure-play cybersecurity company, McAfee, has customers in over 182 countries and supports 78% of Fortune 500 firms. The company is the second largest provider of cybersecurity solutions globally, protecting users all over the world.

McAfee has worked with Welocalize for over five years, developing a Neural Machine Translation-based AI-enabled linguistic process in over 10 languages that pushes the boundaries of NMT in localization to support a global content transformation strategy.

Need for Speed & Accuracy

Speed of software delivery and accessible support materials is a key competitive advantage for technology organizations. There is a continual drive for product updates and knowledge bases, which work to proactively educate and support customers. Any content published for local audiences must keep up with the pace of technology, accurately reflect the brand’s identity, and maintain global consistency in terminology.

Our MT collaboration with Welocalize enables us to localize content at a tailored service level to meet budgetary constraints for different content types. The program continues to be innovative by providing new solutions and thought leadership.” Morgan O’Brien, Technical Program Manager, McAfee

To support the rapid distribution of multilingual content, McAfee utilizes a multi-engine strategy with ongoing customization and training of the MT engines. McAfee combines several NMT engines in their technology stack. In addition to raw MT, the program also combines post-edited MT (MTPE) for knowledge base content to meet the different quality levels required. Post-editing resources are primarily selected based on suitability for the program and knowledge of the product, brand, and experience with relevant content types.

What are the benefits of customized MT Engine Training?

For any enterprise-level MT deployment, there’s a need to customize and train industry and company specific MT engines. Training the engines means they better understand the content and can take into account the context of the translation and specific terminology, leading to more accurate and consistent translations in line with a brand’s style and tone.

Training McAfee’s MT Engines

Welocalize customizes and trains McAfee’s MT engines for each product, serving each content type, which increases accuracy and quality levels enabling more content to be processed. The teams work together, looking at TMs and glossaries, to consolidate, clean, and monitor the performance of the MT engines – matching them with the right content type and language.

Working with Best-in-Class MT Systems

McAfee leverages cutting-edge technologies, working with best-in class MT systems, to fit the specific translation brief, budget and data security requirements. The teams combine and pilot leading commercial solutions including Microsoft Translate, Google AutoML Translation, DeepL Translator as well as Welocalize’s proprietary NMT engines.

Welocalize is an early access partner with AWS, Google, and Microsoft to test and pilot their emerging NMT methodologies.

As an innovative, forward-looking company, McAfee are very much early adopters of the latest MT innovations to drive global strategy. They continually push the boundaries of MT to understand and publish more product and support content in local languages to optimize their security solutions.” Olga Beregovaya, VP, AI Solutions, Welocalize

ICYMI: ISO-Certified PEMT Solutions

Welocalize Life Sciences has gained ISO 18587:2017 certification, which provides requirements for the process of full, human post-editing of machine translation output! As Welocalize’s fourth ISO certification, it represents a quality portfolio found at only 8.57% of LSPs (Source: Slator).