taus-logo-normal
menu-icon
EPIC

OVERVIEW

What is EPIC?

The platform in a nutshell

Choosing a QE Model

Match model to content & risk

EPIC FEATURES

Quality Estimation

Score translations automatically

Automated Post-Editing

Clean up MT output at scale

Retrieval-Augmented Generation

Feed context with glossaries

Specialized Models

Language specific models

USE EPIC

Integrations

Connect with your TMS & Tools

API Docs

Technical docs & API references

EPIC APP

Interactive playground & sandbox
Resources

Blogs

Perspectives from our team

Reports

In-depth industry research

Webinars

Live and on-demand sessions

Case Studies

Real results from real teams
icon epic arrow
icon epic arrow
TAUS Quality Estimation Benchmarking Report
March 2025

This report presents the results of a benchmarking exercise, outlining our testing methodology, the training process, and all the key insights gained from analyzing the correlation between QE scores and human evaluations.

icon epic arrow

Contents of the Report

The Setup - Defining the Benchmarking Framework and Evaluation Process

The Results - Correlation between QE Scores and Human Labels

Key Takeaways - Strengths & Limitations

Case Studies

icon epic arrow
Authors
David Koot
ML Engineer | TAUS
icons-social-media-linked-in-circle
Anne-Maj van der Meer
Head of Sales & Marketing | TAUS
icons-social-media-linked-in-circle