Conversational Music
Recommendation

About

The Music Conversational Recommendation Challenge (Music-CRS) is a global research competition bridging Natural Language Processing and Recommender Systems. Participants build systems that engage in multi-turn conversations about music, recommend relevant tracks, and generate natural language responses.

The challenge uses TalkPlayData-Challenge, a large-scale synthetic multi-turn dialogue dataset grounded in real music listening histories, with pre-extracted multimodal track and user embeddings provided.

Organizing Committee

• Seungheon Doh — KAIST, South Korea
• Sergio Oramas — Pandora/SiriusXM
• Bruno Sguerra — Deezer Research
• Abhinav Bohra — Amazon
• Claudio Pomo — Politecnico di Bari, Italy
• Francesco Barile — Maastricht University, Netherlands

Resources

Conversation Datasets

• Train/Development Dataset: https://huggingface.co/datasets/talkpl-ai/TalkPlayData-Challenge-Dataset
• Blind Dataset A (will be released after the challenge): https://huggingface.co/datasets/talkpl-ai/TalkPlayData-Challenge-Blind-A
• Blind Dataset B (will be released after the challenge): #

Recommendation Environments Datasets

The following resources are shared across all datasets (Train, Development, and Blind sets).

• User Metadata: https://huggingface.co/datasets/talkpl-ai/TalkPlayData-Challenge-User-Metadata
• Track Metadata: https://huggingface.co/datasets/talkpl-ai/TalkPlayData-Challenge-Track-Metadata
• User Embedding: https://huggingface.co/datasets/talkpl-ai/TalkPlayData-Challenge-User-Embeddings
• Track Embedding: https://huggingface.co/datasets/talkpl-ai/TalkPlayData-Challenge-Track-Embeddings

Code

• Baseline Code: https://github.com/nlp4musa/music-crs-baselines
• Evaluation Code: https://github.com/nlp4musa/music-crs-evaluator

Timeline