Driving Evaluation Interview Dataset

This repository contains the post-driving interview dataset and knowledge graphs for the paper:
"A Comprehensive LLM-powered Framework for Driving Intelligence Evaluation"

.
├── knowledge_graph/
│   ├── driver_kg.json         # Knowledge graph for driver interviews
│   └── passenger_kg.json      # Knowledge graph for passenger interviews
│
└── optimised_raw_interview/
    ├── optimised_driver/      # Post-driving interviews (drivers)
    │   └── [audio/text files] # Optimized + translated
    │
    └── optimised_passenger/   # Post-driving interviews (passengers)
        └── [audio/text files] # Optimized + translated

• Knowledge Graphs: Structured representations (JSON) of driver/passenger interview semantics.
• Optimised Interviews: Raw interviews post-processed for:
  • Fluency improvement (LLM-based text optimisation)
  • ASR error correction (speech recognition fixes)
  • English translation (original language: Chinese])