A Simulated Dataset for Proactive Robot Task Inference from Streaming Natural Language Dialogues

Overview

A dataset generated by a large language model, designed to facilitate research on proactive robots that infer human needs from natural language conversations.

Generation

1. Set environment variable OPENAI_API_KEY:

   export OPENAI_API_KEY="your-api-key-here"

2. Generate 10 scenes, items, config files:

   python gen_scene_config.py

3. Generate items, topics; sample streaming topics; generate dialogues and save to dataset:

   python gen_main.py