Uponix, Uponati’s core engine
Uponix is our own large-scale text analysis engine. It uses dependency parsing and named-entity recognition to make large volumes of Korean text machine-workable, and purpose-built variants sit on top of it across our products.
Built on dependency parsing and named-entity recognition
Uponix analyzes large volumes of text into a machine-workable structure using Korean dependency parsing and named-entity recognition. On top of this core capability, we build purpose-built variants that are reused across different products for translation, analysis, and reporting.
What makes up the engine
Dependency parsing
Analyzes sentence structure to correctly handle Korean-specific omission and word-order variation.
Named-entity recognition
Identifies people, organizations and terms to turn documents into structured data.
Purpose-built variants
Specialized models for translation (Uponix-Translate), analysis and reporting sit on top of the core engine.
Backed by patents
Secured as rights through 2 granted patents, including Korean dependency parsing (10-1646461) and named-entity recognition corpus generation (10-1724398).
Where Uponix runs today
KimPro translation
The variant specialized for web-novel translation. It preserves emotional nuance, character voice and series continuity.
Loadly report analysis
Analyzes load-test results and drafts a standard report automatically.
Korean language understanding
A patent-backed AI product that analyzes large volumes of Korean text with dependency parsing and named-entity recognition.
Not sure where to start with digital transformation?
We will assess an adoption plan tailored to your organization.