← Back to explorer

ConvApparel: A Benchmark Dataset and Validation Framework for User Simulators in Conversational Recommenders

Type
paper
Venue
EACL 2026
Year
2026
Source
huggingface
Access
public
Language
en
Added
2026-09-29
Verified
2026-09-29

Summary

Introduces ConvApparel, a dataset of rater-assistant apparel-shopping conversations, plus a validation framework for user simulators used to evaluate conversational recommender systems.

Research notes

  • Key findings: Provides real human shopping dialogues with post-session surveys as ground truth for validating user simulators
  • Paper record for the research behind the ConvApparel dataset; authors not listed on the dataset card.