Towards Robust Version Identification in the Wild: A Dataset, Benchmark, and Fine-Tuning Study

• URL: http://arxiv.org/abs/2608.04543v1
• Date: Wed, 05 Aug 2026 07:28:25 GMT
• Title: Towards Robust Version Identification in the Wild: A Dataset, Benchmark, and Fine-Tuning Study
• Abstract summary: We introduce DiVers, a large-scale VI dataset comprising over 1.1 million musical versions.<n>DiVers provides automatically assigned tags and segment-level predictions.<n>We evaluate the proposed dataset by training state-of-the-art VI systems.
• Score: 9.415756457507383
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Existing datasets for musical version identification (VI) are primarily derived from curated metadata sources such as SecondHandSongs and Discogs, and are therefore dominated by professionally recorded tracks. This leads to a domain mismatch with real-world scenarios, where amateur and user-generated content is prevalent. To address this limitation, we introduce DiVers, a large-scale VI dataset comprising over 1.1 million musical versions, with train-validation-test splits compatible with established datasets such as Discogs-VI-YT, SHS100K, and Da-TACOS. In addition to standard version-level annotations, DiVers provides automatically assigned tags (e.g., instrumental, live) and segment-level predictions indicating the presence or absence of music. We evaluate the proposed dataset by training state-of-the-art VI systems. Our results show that models trained on DiVers achieve substantially improved robustness to acoustically diverse and noisy inputs, while maintaining a stable performance on cleaner, studio-quality benchmarks. We release the dataset metadata, code for its construction, and all experimental pipelines to support reproducibility.

Related papers

• RealDESED: A Real-World Domestic Sound Event Detection Benchmark [41.94606140754546]
  RealDESED is a real-world domestic sound event detection (SED) benchmark comprising 5,710 audio recordings collected by 652 participants in their homes.<n>Each recording is between 15 and 35 seconds long and contains temporally precise annotations for 15 common domestic sound classes.
  arXiv  Detail & Related papers  (2026-07-18T09:50:21Z)
• Scaling Audio-Visual Quality Assessment Dataset via Crowdsourcing [62.250874651622574]
  We propose a practical approach for AVQA dataset construction.<n>We design a crowdsourced subjective experiment framework for AVQA, breaks the constraints of in-lab settings, and achieves reliable annotation across varied environments.<n>We validate this approach through YT-NTU-AVQ, the largest and most diverse AVQA dataset to date, consisting of 1,620 user-generated audio and video sequences.
  arXiv  Detail & Related papers  (2026-02-26T06:18:11Z)
• ConceptCaps -- a Distilled Concept Dataset for Interpretability in Music Models [0.10923877073891443]
  ConceptCaps is a dataset of 23k music-caption-audio triplets with explicit labels from a 200-attribute taxonomy.<n>A VAE learns plausible attribute co-occurrence patterns, a fine-tuned LLM converts attribute lists into professional descriptions, and MusicGen synthesizes corresponding audio.<n>We validate the dataset through audio-text alignment (CLAP), linguistic quality metrics (BERTScore, MAUVE), and TCAV analysis confirming that concept probes recover musically meaningful patterns.
  arXiv  Detail & Related papers  (2026-01-20T17:04:08Z)
• OpenDataArena: A Fair and Open Arena for Benchmarking Post-Training Dataset Value [74.80873109856563]
  OpenDataArena (ODA) is a holistic and open platform designed to benchmark the intrinsic value of post-training data.<n>ODA establishes a comprehensive ecosystem comprising four key pillars: (i) a unified training-evaluation pipeline that ensures fair, open comparisons across diverse models; (ii) a multi-dimensional scoring framework that profiles data quality along tens of distinct axes; and (iii) an interactive data lineage explorer to visualize dataset genealogy and dissect component sources.
  arXiv  Detail & Related papers  (2025-12-16T03:33:24Z)
• Match Stereo Videos via Bidirectional Alignment [15.876953256378224]
  Recent learning-based methods often focus on optimizing performance for independent stereo pairs, leading to temporal inconsistencies in videos. We introduce a novel video processing framework, BiDAStereo, and a plugin stabilizer network, BiDAStabilizer, compatible with general image-based methods. We present a realistic synthetic dataset and benchmark focused on natural scenes, along with a real-world dataset captured by a stereo camera in diverse urban scenes for qualitative evaluation.
  arXiv  Detail & Related papers  (2024-09-30T13:37:29Z)
• SpoofCeleb: Speech Deepfake Detection and SASV In The Wild [76.71096751337888]
  SpoofCeleb is a dataset designed for Speech Deepfake Detection (SDD) and Spoofing-robust Automatic Speaker Verification (SASV)<n>SpoofCeleb comprises over 2.5 million utterances from 1,251 unique speakers, collected under natural, real-world conditions.
  arXiv  Detail & Related papers  (2024-09-18T23:17:02Z)
• Fact Checking Beyond Training Set [64.88575826304024]
  We show that the retriever-reader suffers from performance deterioration when it is trained on labeled data from one domain and used in another domain. We propose an adversarial algorithm to make the retriever component robust against distribution shift. We then construct eight fact checking scenarios from these datasets, and compare our model to a set of strong baseline models.
  arXiv  Detail & Related papers  (2024-03-27T15:15:14Z)
• MARBLE: Music Audio Representation Benchmark for Universal Evaluation [79.25065218663458]
  We introduce the Music Audio Representation Benchmark for universaL Evaluation, termed MARBLE. It aims to provide a benchmark for various Music Information Retrieval (MIR) tasks by defining a comprehensive taxonomy with four hierarchy levels, including acoustic, performance, score, and high-level description. We then establish a unified protocol based on 14 tasks on 8 public-available datasets, providing a fair and standard assessment of representations of all open-sourced pre-trained models developed on music recordings as baselines.
  arXiv  Detail & Related papers  (2023-06-18T12:56:46Z)
• Data leakage in cross-modal retrieval training: A case study [16.18916188804986]
  We study the recently proposed SoundDesc benchmark dataset, which was automatically sourced from the BBC Sound Effects web page. We find that SoundDesc contains several duplicates that cause leakage of training data to the evaluation data. We propose new training, validation, and testing splits for the dataset that we make available online.
  arXiv  Detail & Related papers  (2023-02-23T09:51:03Z)
• Exploring the Efficacy of Pre-trained Checkpoints in Text-to-Music Generation Task [86.72661027591394]
  We generate complete and semantically consistent symbolic music scores from text descriptions. We explore the efficacy of using publicly available checkpoints for natural language processing in the task of text-to-music generation. Our experimental results show that the improvement from using pre-trained checkpoints is statistically significant in terms of BLEU score and edit distance similarity.
  arXiv  Detail & Related papers  (2022-11-21T07:19:17Z)
• Unaligned Supervision For Automatic Music Transcription in The Wild [1.2183405753834562]
  NoteEM is a method for simultaneously training a transcriber and aligning the scores to their corresponding performances. We report SOTA note-level accuracy of the MAPS dataset, and large favorable margins on cross-dataset evaluations.
  arXiv  Detail & Related papers  (2022-04-28T17:31:43Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.