--- title: "Product update: improved translation review categories" description: We have an exciting new update to the Lilt user interface - improved translation review categories. Learn about this new release, and how we're making translator and reviewer workflows more efficient. image: https://labs.lilt.com/hubfs/screen_shot_2019-02-23_at_1.22.36_pm.png --- - [Lilt.com](https://lilt.com/) - [Resources](https://lilt.com/resources) - [Lilt.com](https://lilt.com/) - [Resources](https://lilt.com/resources) ## Product update: improved translation review categories ![](https://labs.lilt.com/hubfs/adrienne.jpeg) by Adrienne Lumb February, 25, 2019 2 Minute Read - - - We’re pleased to announce a new feature in our interface that improves the translation review process. Previously, reviewers could make general text comments about errors and introduce categories by using hashtags to indicate the type of error (such as #punctuation or #mistranslation). However, it was still a manual process to collect the different hashtags and identify precisely where the error occurred. Now, when reviewers find an error in a translation, they can categorize the type of error made - allowing them the ability to provide more detailed feedback. With the new experience, reviewers can select the kind of error (or errors) from a checklist of commonly made mistakes. The various error types are now fields within the revision report document. The reviewer can also now see exact error placement along with the specific error type, providing a birds-eye view over the kinds of errors that occur most frequently in their projects. From here, translation managers can adjust their text specifications, and experience even greater efficiency gains. We love making workflow improvements such as this one, and the structure it provides in feedback loops. Reviewers will now see a set of checkboxes to indicate the type of error found, and will still have the option to leave plain text comments as well. Here’s the new experience: ![](https://labs.lilt.com/hubfs/screen_shot_2019-02-23_at_1.22.36_pm.png) "The addition of translation review categories makes it much easier to see trends in the type of errors," said Translation Manager Phoebe Killick. "Previously, I had to manually look through the delivery revision report to find trends in the types of errors. Now, I can immediately aggregate and filter by types of errors to find individual translators who are consistently making the same errors, or update the instructions for all translators where there are frequent similar errors across the board." We’d love your feedback on the new experience. Let us know your thoughts [@LiltHQ](https://twitter.com/LiltHQ?lang=en) on Twitter. ## Localization is profitable. You can measure it. Having a hard time tying localization to ROI? We've got you. The Ultimate Guide to Measuring Localization ROI covers everything you need to know to tie your market expansion efforts to business goals and revenue. [![Get the guide](https://no-cache.hubspot.com/cta/default/4453601/70b1be9e-a5b7-4bc4-9757-67f96976a1ff.png)](https://cta-redirect.hubspot.com/cta/redirect/4453601/70b1be9e-a5b7-4bc4-9757-67f96976a1ff) [View All Posts](https://labs.lilt.com) ### [August, 2, 2017 What We’re Reading: Neural Machine Translation with Reconstruction 1 Minute Read ![](https://labs.lilt.com/hubfs/spence.jpeg) Neural MT systems generate translations one word at a time. They can still generate fluid translations because they choose each word based on all of the words generated so far. Typically, these systems are just trained to generate the next word correctly, based on all previous words. One systematic problem with this word-by-word approach to training and translating is that the translations are often too short and omit important content. In the paper Neural Machine Translation with Reconstruction, the authors describe a clever new way to train and translate. During training, their system is encouraged not only to generate each next word correctly but also to correctly generate the original source sentence based on the translation that was generated. In this way, the model is rewarded for generating a translation that is sufficient to describe all of the content in the original source. Read More](https://labs.lilt.com/what-were-reading-neural-machine-translation-with-reconstruction?hsLang=en) ### [![word-alignment-machine-translation](https://labs.lilt.com/hubfs/word-alignment.png) July, 6, 2020 A New Age for Word Alignment in Machine Translation 4 Minute Read ![](https://labs.lilt.com/hs-fs/hubfs/Drew%20Evans_Edited%20Headshots-large-0725%20copy.jpg?width=299&quality=low) Here at Lilt, we have a team full of exceptionally smart and talented individuals that are working hard to solve the translation and localization industries’ toughest challenges. We’re always researching new ways to improve the day-to-day lives of localization leaders and translators alike. Read More](https://labs.lilt.com/a-new-age-for-word-alignment-in-machine-translation?hsLang=en) - [System Status](https://status.lilt.com/) - [Support Terms](https://lilt.com/terms) - [Privacy](https://lilt.com/privacy) - - - -  2026 Lilt, Inc.