Getting Started with Lakra

Welcome to Lakra! This guide will help you get started with the annotation and evaluation system for machine translation quality assessment.

What is Lakra?

Lakra is a comprehensive annotation and evaluation platform designed specifically for the WiMarka project. It combines AI-powered assessment with human-in-the-loop evaluation to provide comprehensive quality analysis for machine translation.

Key Capabilities

  • Interactive Annotation: Highlight text, classify errors, and provide quality scores

  • AI-Assisted Evaluation: Get automated quality assessments with confidence levels

  • Multi-Role Support: Different interfaces for Annotators, Evaluators, and Administrators

  • Voice Recording: Record audio explanations for corrections

  • Real-time Feedback: Immediate validation and progress tracking

Overview for Different Roles

Annotators

As an annotator, you will:

  • Review source text and machine translations

  • Highlight and classify translation errors

  • Provide quality ratings (fluency, adequacy, overall quality)

  • Suggest corrections and improvements

  • Optionally record voice explanations

Next Steps: See the Annotator Guide for detailed instructions.

Evaluators

As an evaluator, you will:

  • Review annotations created by other users

  • Assess machine translation quality with AI assistance

  • Provide feedback on annotation quality

  • Track evaluation metrics and agreement rates

Next Steps: See the Evaluator Guide for detailed instructions.

Administrators

As an administrator, you will:

  • Manage user accounts and permissions

  • Add and manage sentences for annotation

  • Monitor system statistics and quality metrics

  • Import content in bulk via CSV files

  • Control quality and system settings

Next Steps: See the Admin Guide for detailed instructions.

System Requirements

For End Users

  • Web Browser: Modern browser (Chrome, Firefox, Safari, Edge)

    • JavaScript enabled

    • Cookies enabled for authentication

  • Internet Connection: Stable internet connection required

  • Audio Recording (optional): Microphone for voice annotations

Display Requirements

  • Screen Resolution: Minimum 1280x720 recommended

  • Mobile Support: Responsive design works on tablets and phones

  • Accessibility: Keyboard navigation and screen reader support

Quick Start

  1. Access Lakra: Navigate to your Lakra instance URL

  2. Sign In: Use your provided credentials (email or username)

  3. Complete Onboarding: If required, complete the onboarding test for your role

  4. Start Working: Access your role-specific dashboard

First Steps by Role

For Annotators

Note

After logging in, you’ll see the annotation dashboard with available sentences.

  1. Log in with your credentials

  2. Complete onboarding test (if required)

  3. Navigate to “Start Annotation”

  4. Select a sentence to begin

  5. Follow the annotation workflow

For Evaluators

  1. Log in with your credentials

  2. Complete onboarding test (if required)

  3. Access the evaluation dashboard

  4. Review pending evaluations or quality assessments

  5. Provide feedback on annotations

For Administrators

  1. Log in with administrator credentials

  2. Access the admin dashboard

  3. Review system statistics

  4. Set up users and content as needed

  5. Monitor quality metrics

Getting Help

  • In-App Guidelines: Click the info icon (ℹ️) for contextual help

  • FAQ: See the FAQ section for common questions

  • Support: Contact your system administrator

What’s Next?