A Semantically Consistent Dataset for Data-Efficient Query-Based Universal Sound Separation was created by researchers from Tsinghua University, Shanda AI, and Johns Hopkins University. The dataset is designed for universal sound separation tasks using query-based methods. It was published on Arxiv in 2026 and is hosted on Hugging Face by the author JusperLee.
Use Cases
- Training query-based universal sound separation models based on the described semantic consistency.
- Benchmarking data-efficient audio separation algorithms based on the dataset's design.
- Researching semantic audio representations for separation tasks based on the dataset's focus.
Strengths
- Designed for semantic consistency, a specific feature mentioned in the description.
- Created by a collaboration of institutions including Tsinghua University and Johns Hopkins University.
- Associated with a 2026 Arxiv publication, indicating recent research activity.
Limitations
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count is unknown, which may limit suitability assessment.
- Description metadata is limited; actual data quality requires manual inspection after download.
Provenance
- Source
- Tsinghua University, Shanda AI, Johns Hopkins University
- Freshness
- Last updated 2026-05-19 15:44:55; freshness should be verified.