current landscape & challenges
Translation of global scientific documents is a key requirement across several Regulatory/R&D functions ranging from Labeling, Artwork to local multi-lingual literature monitoring. This has traditionally been performed by local language experts, language translation companies, regional affiliates, etc. and has been a costly and complex global coordination function, covering multiple time zones.
our automation solution
Freyr has approached to solve the above challenges by taking a 3-step approach to automation of scientific/medical translation.
Step 1: Identify a stage-gate approach to translation, rather than a holistic translation even for simple, structured search terms in target languages.
Step 2: Once triaged for initial stage-gate-based simple keywords, identify contextual content/expanded search criteria.
Step 3: Once the baseline search criteria are set, the use of NLP-based translation tools like Amazon Translate, Microsoft Translator, etc. in conjunction with contextual NLP models using Amazon Comprehend/Amazon Medical Comprehend, etc. can significantly reduce translation cost and timelines.
OptPost auto-translation has a linguistic expert to QC the document if required.
Option to add an ML model to learn corrections done by a linguistic expert to improve medium to long-term translation accuracy.
real-world benefits
- Exponential reduction in translation costs.
- Significantly improved centralized translation process.
- Less dependence of expensive, local language / scientific experts – especially for simple translation which can be automated.
