Machine Translation Engine
Saltlux’s Ai Suite machine translation engine is based on an artificial neural network, which has evolved from Saltlux’s existing statistical machine translation engine. It is also integrated and configured with CAT (Computer-assisted translation) tools to specialize in machine translation learning for each domain to improve translation quality and productivity.
< Artificial neural network machine translation engine concept map >
Main Features
- Ease of domain application
The system allows machine translation models to be created for each customer’s specific field, using specialized data processing tools in line with customer data. The domain-specific machine translation model is combined with the general model to form a new model that enables good translation in the customer’s domain and general field knowledge inheritance.
- Quality improvement through automatic translation post-processing and re-learning
The continuous cycle of the machine translation system includes NMT (Neural Machine Translation) learning, translation application, quality feedback, and automatic relearning to improve translation quality. With such a design, the machine translation system is increasingly intelligent, allowing higher quality the longer you use the translation tool.
< Machine translation engine’s virtuous cycle work structure >
- Connectivity to various office tools
The machine translation engine provides an Office plug-in, which enables integration with office tools. Through this, machine translation results can be immediately used in different Office programs.
Main features and specifications
The Machine translation engine consists of a data area for searching and managing translation resources and dictionaries, NMT area for performing machine translation, and management tool area for managing translation results.
< Machine translation engine feature block diagram >
- Machine translation learning and model management
It includes learning model management and evaluation tools for machine translation quality control. The learning model management function may build a variety of machine translation models by changing the learning data and adjusting the parameters, as well as support evaluation and comparative analysis functions for each generated model.
- Learning data and dictionary management
It includes functions for managing translation dictionaries and learning translation data, allowing you to apply machine translation to the domain and better manage translation quality in that domain.
- Service management feature
It offers machine translation service management capabilities such as API service management, machine user management, and machine translation service usage status search.
- Translation result confirmation and post-translation editing management
This function allows you to revise translation results. Translation changes are reflected in the learning model, influencing the following translation. The more you use the system through this virtuous cycle of learning, translation, monitoring, and re-learning, the better the quality of translation can be.