Wednesday 6 November 2024
Universal Language AI's highly-praised Lingo Bagel performs brilliantly in translating financial reports, medical documents, and games
To meet the strong market demand for translation tools, translation technology developments have been well underway for years. However, machine translation of natural languages has mostly been based on statistical models or linguistic rules coupled with vocabulary and sentence correlations built from a large parallel corpus. These types of tools are mostly for general-purpose use and are only good for translating commonly used terms or simple everyday conversations. They are inadequate for professional applications such as translating books and financial reports. These types of translation tasks often require qualified human translators but it may take longer and cost more.According to Universal Language AI COO Yu-De Fei (Alan), a financial report is generally longer than 200 pages and has more than 200,000 words. It takes months to translate a financial report. In the case of a tight deadline, the company will need to hire multiple professional translators and pay a higher fee to have its financial report translated in time. When it comes to translation for medical, financial, mechanical, and legal sectors, translators with field-specific knowledge are needed. Not only are they hard to find but they are also expensive to hire. Many companies are endeavoring to use generative AI to develop automated translation solutions. Lingo Bagel focuses on the translation of financial reports and medical documents. It collaborates with multiple translation agencies to allow the expertise of professional translators to deliver maximum benefit. Lingo Bagel also builds dedicated translation models for companies to guarantee top-quality and top-speed translation services.Using a dedicated model coupled with a professional terminology database, Lingo Bagel completes financial report translation in 60 minutesWith traditional translation, the process takes a long time, the quality may be poor and it is difficult to find professional native speakers. To address the three major pain points, Universal Language AI, established in 2023, used AI coupled with a group of accountants' expertise to develop Lingo Bagel. Its solution has been successfully used by a great many customers. First of all, Universal Language AI worked with dozens of accountants to build a professional terminology database containing more than 2,000 terms compliant with the International Financial Reporting Standards (IFRS). Secondly, it built a dedicated AI model for financial reports, which together with the professional terminology database ensures the terms used in the translation are correct and consistent. To speed up the translation process, Universal Language AI incorporated a systematic workflow, which enables Lingo Bagel to complete the translation of a 200-page, 200,000-word financial report in 60 minutes. Alan noted that Lingo Bagel can precisely translate Chinese financial reports containing professional finance and accounting terminologies into English even if they come in different structures. Lingo Bagel has helped multiple small and mid-size accounting firms complete the translation of financial reports totaling more than one million words and one thousand pages. With respect to translation speed, Lingo Bagel completed the translation of a nursing book that has 400,000 words and 800 pages into Chinese in five days for Taipei Medical University. According to a translation agency that the university approached for a quote, the same task would cost NT$1 million and take 18 months. With Lingo Bagel, the cost was reduced to just one-tenth and the amount of time saved was far more significant. Apart from financial reports and medical books, Universal Language AI has also expanded into game and press release translation. For example, Lingo Bagel translated a Japanese game for a gaming company. The translated script was given back to the game marketing team in a snap for proofreading and polishing, significantly shortening the upgrade cycle. Lingo Bagel also translated a Korean press release for a gaming company. The task was assigned in the morning and completed in the afternoon on the same day. The amazing speed largely boosted the efficiency of press release distribution.Leveraging Startup Terrace Kaohsiung AWS JIC resources to target global marketsAs a seasoned entrepreneur, Alan learned of the Startup Terrace Kaohsiung AWS JIC program and immediately pushed for Universal Language AI to participate in the program. He hopes that the startup resources made available through the project will accelerate Universal Language AI's growth and help expand its footprint in the global marketplace.According to Alan, Lingo Bagel uses Amazon Rekognition from Amazon Web Services (AWS), which automates image recognition and video analysis, to quickly identify and extract the text on the customer's user interface and perform comparisons across different languages. On top of that, it uses Amazon Bedrock, a fully managed service that makes high-performing foundation models (FMs) from leading AI companies, to generate UI text and polish translation results, thereby optimizing Lingo Bagel's service quality and performance. Lingo Bagel currently offers translation in multiple languages, including Chinese-English, Chinese-Korean, Chinese-Japanese, Korean-English, and Japanese-English two-way translation services. Universal Language AI is reaching out to translation agencies across the globe for partnership opportunities.In addition, Universal Language AI plans to work with publishing houses in Taiwan to translate more foreign language books into Chinese, allowing people in Taiwan to read more quality books.