Pith. sign in

REVIEW 3 major objections 9 minor 38 references

On the Robustness of Cover Version Identification Models: A Study Using Cover Versions from YouTube

T0 review · 3 major / 9 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The paper claims that state-of-the-art cover song identification models, which score well on community-built SecondHandSongs benchmarks, lose substantial ranking accuracy on YouTube-style cover versions, especially when the version…

desk verdict Good dataset and taxonomy, but the headline robustness claim is confounded by the uncertainty-sampling design; the paper deserves review with revisions. read the letter →

arxiv 2501.01333 v1 pith:SJWHFJ5A submitted 2025-01-02 cs.MM cs.IRcs.SI

Simon Hachmeier , Robert Jäschke This is my paper · ORCID
classification cs.MMcs.IRcs.SI
keywords coversongidentificationversionYouTubebenchmarkdatasetuncertaintysamplingrobustnessmusicinformationretrievaltaxonomyofalterations
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper asks whether cover song identification models, which are trained and evaluated on collections derived from the community database SecondHandSongs, transfer to cover versions as they actually appear on YouTube. To answer it, the authors build SHS-YT, a 900-video benchmark selected by a multi-modal uncertainty sampling procedure and labelled by crowd workers and music experts, then evaluate several state-of-the-art models on it. They find a large performance drop: CoverHunter's mean average precision falls from 0.86 on SHS100K-Test to 0.52 on SHS-YT+2Q. The paper also identifies which alteration types are hardest—instrumental, drum-only, and vocal-only versions, plus medleys and videos mixing music and speech—and offers a taxonomy of alterations in online-video covers. If correct, the result means current benchmarks overstate how robust cover song identification is in the wild.

What carries the argument

The load-bearing mechanism is the multi-modal uncertainty sampling pipeline used to build the dataset. For each candidate YouTube video, an audio proxy (pre-trained Re-MOVE, which embeds CREMA features into 256-dimensional vectors) computes mean cosine similarity against all seed versions of a work, while a text proxy (Ditto, a BERT-based entity matcher) computes a maximum matching confidence from titles and channel names. Candidates are ranked by disagreement between the two proxies (audio-high/text-low and text-high/audio-low) and by mutual uncertainty, defined as negative Euclidean distance to the per-work center of uncertainty in the two-dimensional similarity–confidence space. This procedure is designed to over-sample out-of-distribution candidates, i.e., videos whose relatedness the audio model and the text model cannot agree on. The benchmark then combines the 900 annotated candidates with query versions from SHS-SEED to form SHS-YT+2Q and SHS-YT+AllQ, and evaluation is done with MIREX metrics MAP and MR1.

What would settle it

Take the same 100 works from SHS-SEED, retrieve a random sample of YouTube videos without uncertainty sampling, annotate them with the same ordinal labels, and measure CoverHunter's MAP on the resulting benchmark; if MAP approaches the 0.86 seen on SHS100K-Test, the drop reported here is an artifact of hard-case sampling rather than a property of YouTube covers.

Watch

Extended reading notes

Core claim

The central discovery is that cover version identification models are substantially less robust on YouTube-style covers than standard benchmarks suggest, and that the gap is driven by identifiable alteration types. Using 100 western-popular-music works from SHS100K-Test as seed, the authors retrieve 94,358 candidate YouTube videos, select 900 candidates with disagreement and mutual-uncertainty sampling between an audio similarity proxy (Re-MOVE embeddings) and a text matching proxy (Ditto), and annotate them on an ordinal relevance scale (NoMusic, NonVersion, Version, Match). On the resulting benchmark SHS-YT+2Q, all tested models rank worse than on SHS100K-Test; CoverHunter drops from 0.86 to 0.52 MAP, CQTNet from 0.66 to 0.50, Re-MOVE from 0.53 to 0.40. Similarity analysis shows that positive pairs from YouTube are significantly less similar than SHS positive pairs, and that non-versions from YouTube are more similar than random negative pairs, which explains the ranking degradation. The paper's taxonomy classifies alterations into Song-level attributes (melody, harmony, timbre, key, timing, tempo) and Video-level context (multiple songs, non-music noise, chunking, fidelity, stem isolation, in-background), with stem isolation and video context as the hardest cases.

Load-bearing premise

The whole comparison rests on the assumption that the uncertainty-sampled candidates represent the distribution of cover versions on YouTube; if the sampling over-selects hard or unusual cases, the measured performance drop may not generalize to YouTube covers at large.

Editorial extensions

If this is right

  • If the gap is real, published MAP numbers on SHS100K-Test and Da-Tacos overstate real-world robustness on YouTube-like content, especially for copyright-detection and recommendation applications.
  • Instrumental, drum-only, vocal-only, and single-instrument versions are consistently hard for current models, pointing to the absence of the main melody and harmony as a key failure cause.
  • Medleys, videos with multiple versions, and videos with non-musical speech break the alignment assumptions baked into current models, so chunk-based matching and noise-mixing augmentation are natural next steps.
  • The ordinal labels (NoMusic, NonVersion, Version, Match) and 19 uncertainty classes give a finer-grained evaluation target than binary same-work labels.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The uncertainty sampling deliberately enriches for hard cases, so the 0.52 MAP on SHS-YT+2Q is likely an upper bound on the difficulty of typical YouTube covers; a random sample of YouTube covers for the same works would probably show a smaller but still positive gap.
  • If stem-isolated versions (drum-only, vocal-only) are the hardest class, then integrating sound-source separation into training augmentation, or evaluating query-by-humming systems on the same SHS-YT candidates, could be a cheap way to improve robustness without new annotation.
  • The taxonomy suggests a testable prediction: models trained with synthetic concatenations of songs plus speech noise, or with sliding-window matching, should reduce the performance gap specifically for the Video: Multiple Songs and Video: With Non-Music classes.
  • Because YouTube is dynamic, the benchmark's URLs may decay; the released CQT and CREMA features are the durable artifact, so future comparisons should rely on those features rather than re-downloading audio.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 9 minor

Summary. The paper asks whether version identification (VI) models validated on SecondHandSongs-based benchmarks generalize to cover versions as they actually appear on YouTube. The authors build a new dataset, SHS-YT: from the first 100 works of SHS100K-Test (SHS-SEED), they retrieve 94,358 YouTube candidates by text queries, then select nine candidates per work via multi-modal uncertainty sampling using Re-MOVE (audio) and Ditto (text), forming three groups (DisagrAudio, DisagrText, MutualUnc). The resulting 900 candidates are labeled by MTurk workers on an ordinal scale (NoMusic, NonVersion, Version, Match) and curated by two experts who add 19 uncertainty classes and a taxonomy of alterations. The authors benchmark CoverHunter, CQTNet, Re-MOVE, Ditto, and a fuzzy baseline on SHS-YT+2Q and SHS-YT+AllQ, finding a strong MAP drop on SHS-YT+2Q (CoverHunter 0.52 vs 0.86 on SHS100K-Test) but a much smaller gap at SHS-YT+AllQ (0.83 vs 0.86), where CQTNet and Re-MOVE even improve on their SHS100K-Test scores. Cosine-similarity analyses (Tables 6-7) show that YouTube positives are less similar and YouTube negatives more similar than their SHS counterparts, with drum-only, instrumental, medley, and vocal-only versions being the hardest classes.

Significance. The paper has genuine value even though the headline drop is not a clean measurement of general robustness on YouTube. The dataset's ordinal labels (Match vs Version vs NonVersion vs NoMusic), the expert uncertainty classes, and the alteration taxonomy are reusable resources; the repository ships URLs, queries, and CQT/CREMA features, making most quantitative claims checkable; and the falsifiable finding that drum-only, instrumental, medley, and vocal-only versions are hardest for all three audio models (Table 7) is directly actionable for training-augmentation research. The paper is also partially transparent about its own confound: it reports SHS-YT+AllQ where the drop shrinks to 0.83 vs 0.86, and it flags the Re-MOVE/Ditto selection bias in a closing remark. The significance is conditional: the abstract-level claim of significantly lower ranking performance is not yet separated from dataset-construction effects, so the current evidence supports a narrower claim about uncertainty-selected hard cases rather than a general robustness deficit of VI models on YouTube.

major comments (3)
  1. [Abstract; §5.1 (Table 5); §6] The headline comparison, CoverHunter MAP 0.86 on SHS100K-Test vs 0.52 on SHS-YT+2Q, is not like-for-like, and the paper draws a broader conclusion from it than the design supports. SHS-YT+2Q is composed by construction of the nine most uncertain candidates per work (Section 3.2), and its negatives come largely from the same YouTube search result sets: Table 6 shows that both construction effects move the metrics, with YT-Positives less similar to SHS-SEED queries than SHS-Positives (0.80 vs 0.88 for CoverHunter) and YT-Negatives more similar than SHS-Negatives (0.72 vs 0.68). The AllQ column in the same table is decisive: once SHS-SEED versions are added back, CoverHunter's MAP is 0.83 (near the 0.86 baseline) and CQTNet and Re-MOVE score higher than on SHS100K-Test (0.72 vs 0.66 and 0.56 vs 0.53). The claims 'existing models achieve significantly lower ranking performance on our dataset' (Abstract) and 'we in fact confirmed a significant difference' (§6) should therefore be recast as statements about the uncertainty-sampled regime. I recommend adding a control evaluation of the same models on a random, non-uncertainty-filtered subsample of YT-CRAWL, and reporting results per sampling group; the per-group positive rates already differ sharply (93/293, 82/300, 26/307 for DisagrAudio, DisagrText, MutualUnc in Table 4), so aggregate MAP hides which part of the dataset drives the drop.
  2. [§5.1 closing remark; §5.2 (Tables 6-7)] The sampling-bias caveat is confined to Re-MOVE and Ditto, but the bias is a property of the dataset rather than of the proxy models: because every candidate in SHS-YT was selected for proxy disagreement or mutual uncertainty, CoverHunter and CQTNet are evaluated on the same preselected hard set, and the aggregate numbers cannot separate the selection effect from model-level robustness. The issue is compounded for Re-MOVE itself: since DisagrText selects candidates with low Re-MOVE audio similarity, Re-MOVE's YT-Positive row in Table 7 (0.45 vs 0.62 for SHS-Positive) is partly by construction. Reporting the similarity statistics separately for each sampling group, or on a random control subset, would let the reader quantify the by-construction component; at minimum, the closing remark in §5.1 should be extended to state that the selection affects the evaluation of all benchmarked models, not only the modality proxies.
  3. [§5.1; Abstract] No uncertainty is attached to the MAP/MR1 point estimates, yet the Abstract's 'significantly lower' is a statistical claim. With only 100 works, bootstrap confidence intervals over query works (or per-work AP distributions) are needed to judge whether the 0.52 vs 0.86 gap, and especially the 0.83 vs 0.86 AllQ gap, is distinguishable from noise. Relatedly, the two-sample t-tests in Tables 6-7 use supports up to 5.6 million pairs, so p<0.01 is reported for mean differences as small as 0.04 (YT-Negative 0.72 vs SHS-Negative 0.68 for CoverHunter); effect sizes or Cohen's d should be reported alongside, and the YT-Match row of Table 6 rests on only 44 pairs.
minor comments (9)
  1. [§5.1] Typo: 'experiements' should be 'experiments'.
  2. [§4.1] Figure 3 caption: 'the SSHS-YT dataset' and 'bandwith' should read 'SHS-YT' and 'bandwidth'.
  3. [§5] In the description of SHS-YT+2Q, 'one additional work from SHS-SEED' should read 'one additional version from SHS-SEED', since each work already contributes two query versions.
  4. [§5.2] Table 7 caption: the baseline is given as '(SHS-Positive or SHS-Positive respectively)'; it should be 'SHS-Positive or SHS-Negative respectively', since the lower block is contrasted with SHS-Negative.
  5. [§4.1] Twenty SHS-YT candidates also appear in SHS100K (13 from training, 2 from validation, 5 from test), all but one labeled NonVersion; since Section 5 excludes only SHS-SEED versions from the benchmark, please clarify whether the single non-NonVersion candidate drawn from the training or validation subset is retained in SHS-YT+2Q and SHS-YT+AllQ, which would constitute a mild training contamination for CoverHunter and CQTNet.
  6. [§3.3] The sentence 'We publish one human intelligence tasks (HITs) on MTurk per work' should read 'one Human Intelligence Task (HIT)'.
  7. [References] Airoldi, Beraldo, and Gandini (2016) is listed twice as references [3] and [4] with identical bibliographic data; one entry should be removed.
  8. [§3.1] The paper does not explain how 'the first 100 works' from SHS100K-Test are ordered; if the repository listing order is arbitrary the choice is acceptable, but the ordering criterion should be stated so that the seed selection cannot introduce an unexamined bias.
  9. [Figure 5] Figure 5 appears to show the same similarity matrix twice and lacks a legend and explicit axis labels for the uncertainty classes; please clarify the layout so that the upper and lower halves are interpretable.

Circularity Check

1 steps flagged · score 2.0 of 10

Minor self-selection bias for proxy models Re-MOVE and Ditto; the central CoverHunter/CQTNet benchmark remains independent.

  1. fitted input called prediction [Section 3.2 (Uncertainty Sampling) and Section 5.1 (Overall Performance, Table 5)]
    "We use the pre-trained model Re-MOVE [37] as a proxy in the audio/music domain which is one of the best approaches for VI at the time of dataset creation. ... select the three candidates with the highest disagreement for both disagreement groups per work. ... A closing remark on the overall evaluation is the potential influence of sampling bias to the performance of Re-MOVE and Ditto, since these models are used as modality proxies in dataset creation."

    Re-MOVE and Ditto are the two proxies that define the uncertainty-sampled test set: candidates are selected by maximal disagreement between or mutual uncertainty of these two models. Benchmarking the same two models on the resulting dataset (Table 5: Re-MOVE MAP 0.40, Ditto MAP 0.39 on SHS-YT+2Q) therefore measures their performance on examples selected specifically to be hard for them; their scores are not independent evidence of out-of-distribution robustness. The paper explicitly acknowledges this bias, so it is not hidden, but the Re-MOVE/Ditto portion of the benchmark is still partly forced by the selection rule. The central claim does not rest on these two models alone, because CoverHunter and CQTNet were not involved in candidate selection and also drop.

full rationale

The paper is an empirical benchmarking study rather than a derivation from first principles, so the central claim is a measurement. The only load-bearing circular ingredient is that Re-MOVE (and Ditto) are used in the uncertainty-sampling step that defines SHS-YT and are then themselves evaluated on the resulting set; their measured MAP drop is partly a consequence of selecting candidates with maximal disagreement or mutual uncertainty for those exact models. The paper acknowledges this in Section 5.1. However, the headline result is based on CoverHunter and CQTNet, which were not used to select the dataset, and those models also show substantial drops (e.g., CoverHunter MAP 0.52 on SHS-YT+2Q vs 0.86 on SHS100K-Test). The cosine-similarity distribution analysis is likewise independent of the proxy selection. The only other self-citation, the query-formulation method of Hachmeier et al., is not load-bearing for the main conclusion. Therefore the circularity is minor and localized to the proxy models' own benchmark numbers, not to the paper's central robustness finding.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

The paper's central claim rests on the reliability of human annotations, the representativeness of the seed works, and the use of embedding cosine similarity as a proxy for ranking difficulty. No new physical or mathematical entities are introduced.

assumptions (3)
  • domain assumption MTurk labels plus expert curation are a reliable ground truth for version relevance.
    Section 3.3: worker agreement is moderate (Krippendorff's alpha 0.43), expert-worker agreement is high (Kendall's tau 0.81); the benchmark depends on this labeling.
  • domain assumption The first 100 works of SHS100K-Test are a representative sample of western popular music on YouTube.
    Section 3: used as the seed set; no evidence that these 100 works cover the diversity of YouTube versions.
  • domain assumption Cosine similarity between Re-MOVE, CQTNet, and CoverHunter embeddings reflects ranking performance.
    Section 5.2: distribution analysis relies on cosine similarities as proxies for retrieval difficulty.

how reviews work

0 comments
Cite this review

Pith. "Pith review of On the Robustness of Cover Version Identification Models: A Study Using Cover Versions from YouTube." pith.science (2026). https://pith.science/paper/SJWHFJ5A

@misc{pith2026250101333,
  author       = {Pith},
  title        = {Pith review of: On the Robustness of Cover Version Identification Models: A Study Using Cover Versions from YouTube},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SJWHFJ5A}},
  note         = {Machine review of arXiv:2501.01333}
}
read the original abstract

Recent advances in cover song identification have shown great success. However, models are usually tested on a fixed set of datasets which are relying on the online cover song database SecondHandSongs. It is unclear how well models perform on cover songs on online video platforms, which might exhibit alterations that are not expected. In this paper, we annotate a subset of songs from YouTube sampled by a multi-modal uncertainty sampling approach and evaluate state-of-the-art models. We find that existing models achieve significantly lower ranking performance on our dataset compared to a community dataset. We additionally measure the performance of different types of versions (e.g., instrumental versions) and find several types that are particularly hard to rank. Lastly, we provide a taxonomy of alterations in cover versions on the web.

Figures

Figures reproduced from arXiv: 2501.01333 by the authors.

Figure 1
Figure 1. Dataset creation. 3.1 Candidate Retrieval The goal of the candidate retrieval step is to obtain a set of candidate versions to be included in our dataset. We apply an approach by Hachmeier et al. [18] to formulate multiple text queries per work in SHS-SEED. We utilize the strings for performer and title of the first version of each work to formulate queries and additionally formulate new queries using YouTube search… view at source ↗
Figure 2
Figure 2. Our instructions and examples to workers as presented on MTurk. Please [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗
Figure 3
Figure 3. We observe a strong peak in duration around 3.5 minutes and in [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figures from the paper (4 more)
Figure 3
Figure 3. Figure 3: Gaussian kernel density estimates for properties of the videos in the SSHS [PITH_FULL_IMAGE:figures/full_fig_p011_3.png]
Figure 4
Figure 4. Figure 4: Relative proportion of uncertainty class annotated. [PITH_FULL_IMAGE:figures/full_fig_p012_4.png]
Figure 5
Figure 5. Figure 5: Mean Cosine similarities of CoverHunter embeddings between YT [PITH_FULL_IMAGE:figures/full_fig_p017_5.png]
Figure 6
Figure 6. Figure 6: Taxonomy of Cover Versions in Online Videos [PITH_FULL_IMAGE:figures/full_fig_p018_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

38 extracted references · 26 canonical work pages

  1. [1]

    Abrassart, M., Doras, G.: And what if two musical versions don’t share melody, harmony, rhythm, or lyrics ? In: International Society for Music Information Re- trieval Conference (2022)

    work page 2022

  2. [2]

    In: Bhatnagar, V., Srinivasa, S

    Agrawal, S., Sureka, A.: Copyright infringement detection of music videos on YouTube by mining video and uploader meta-data. In: Bhatnagar, V., Srinivasa, S. (eds.) Big Data Analytics. pp. 48–67. Springer International Publishing, Cham (2013)

    work page 2013

  3. [4]

    Poetics57, 1–13 (2016).https://doi.org/10

    Airoldi, M., Beraldo, D., Gandini, A.: Follow the algorithm: An exploratory in- vestigation of music on YouTube. Poetics57, 1–13 (2016).https://doi.org/10. 1016/j.poetic.2016.05.001

    work page 2016

  4. [5]

    In: Proceedings of the 25th Interna- tional Society for Music Information Retrieval Conference (ISMIR)

    Araz, R.O., Serra, X., Bogdanov, D.: Discogs-VI: A musical version identification dataset based on public editorial metadata. In: Proceedings of the 25th Interna- tional Society for Music Information Retrieval Conference (ISMIR). San Francisco, CA, USA (2024)

    work page 2024

  5. [6]

    https://doi.org/10.5281/zenodo

    Bachmann, M.: Rapidfuzz: Release 1.8.0. https://doi.org/10.5281/zenodo. 5528870 (2021), zenodo. DOI: 10.5281/zenodo.5528870 20 Simon Hachmeier and Robert Jäschke

    doi:10.5281/zenodo 2021

  6. [7]

    PsyArXiv preprint (2020)

    Daikoku, H., Ding, S., Sanne, U.S., Benetos, E., Wood, A.L., Fujii, S., Savage, P.E.: Human and automated judgements of musical similarity in a global sample. PsyArXiv preprint (2020)

    work page 2020

  7. [8]

    arXiv preprint arXiv:1810.04805 (2018)

    Devlin,J.,Chang,M.W.,Lee,K.,Toutanova,K.:Bert:Pre-trainingofdeepbidirec- tional transformers for language understanding. arXiv preprint arXiv:1810.04805 (2018)

    arXiv 2018

  8. [9]

    In: Inter- national Conference on Acoustics, Speech and Signal Processing (ICASSP)

    Du, X., Chen, K., Wang, Z., Zhu, B., Ma, Z.: Bytecover2: Towards dimensional- ity reduction of latent embedding for efficient cover song identification. In: Inter- national Conference on Acoustics, Speech and Signal Processing (ICASSP). pp. 616–620. IEEE (2022)

    work page 2022

Show all 38 references
  1. [10]

    In: ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

    Du, X., Wang, Z., Liang, X., Liang, H., Zhu, B., Ma, Z.: Bytecover3: Accurate cover song identification on short queries. In: ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). pp. 1–5 (2023). https://doi.org/10.1109/ICASSP493...

    2023

  2. [11]

    International Conference on Acoustics, Speech and Signal Processing (ICASSP) pp

    Du, X., Yu, Z., Zhu, B., Chen, X., Ma, Z.: Bytecover: Cover song identification via multi-loss training. International Conference on Acoustics, Speech and Signal Processing (ICASSP) pp. 551–555 (2020)

    2020

  3. [12]

    In: International Conference on Acoustics, Speech and Signal Processing (ICASSP)

    Ellis, D.P., Poliner, G.E.: Identifying ‘cover songs’ with chroma features and dy- namic programming beat tracking. In: International Conference on Acoustics, Speech and Signal Processing (ICASSP). vol. 4, pp. 1429–1432. IEEE (2007). https://doi.org/10.1109/ICASSP.2007.367348

    2007

  4. [13]

    Journal of New Music Research45, 239 – 251 (2016)

    Flexer, A., Grill, T.: The problem of limited inter-rater agreement in modelling music similarity. Journal of New Music Research45, 239 – 251 (2016). https: //doi.org/10.1080/09298215.2016.1200631

    2016

  5. [14]

    Flexer, A., Lallai, T.: Can we increase inter-and intra-rater agreement in mod- eling general music similarity? In: Conference of International Society for Music Information Retrieval (ISMIR). pp. 494–500 (2019)

    2019

  6. [15]

    Transactions of the International Society for Music In- formation Retrieval4(1), 182–194 (2021).https://doi.org/10.5334/tismir.107

    Flexer, A., Lallai, T., Rašl, K.: On evaluation of inter- and intra-rater agreement in music recommendation. Transactions of the International Society for Music In- formation Retrieval4(1), 182–194 (2021).https://doi.org/10.5334/tismir.107

    2021 doi

  7. [16]

    SSRN (2019)

    Ghosh,S.,Sperling,R.,Hooper,S.:UsingAmazonMTurkforresearchinacademia: A beginner’s guide for using Qualtrics, detecting VPN/proxy, limiting countries using geolocation & other tips. SSRN (2019). https://doi.org/10.2139/ssrn. 3455722

    2019 doi

  8. [17]

    arXiv preprint arXiv:2005.08100 (2020)

    Gulati, A., Qin, J., Chiu, C.C., Parmar, N., Zhang, Y., Yu, J., Han, W., Wang, S., Zhang, Z., Wu, Y., et al.: Conformer: Convolution-augmented transformer for speech recognition. arXiv preprint arXiv:2005.08100 (2020)

    2020 arXiv

  9. [18]

    (eds.) Proceedings of the LWDA 2021 Work- shops: FGWM, FGKD, and FGDB, Hildesheim (Germany), Oktober 5-7th, 2022

    Hachmeier, S., Jäschke, R., Saadatdoorabi, H.: Music version retrieval from youtube: How to formulate effective search queries? In: Reuss, P., Eisenstadt, V., Schönborn, J.M., Schäfer, J. (eds.) Proceedings of the LWDA 2021 Work- shops: FGWM, FGKD, and FGDB, Hildesheim (German...

    2022

  10. [19]

    Contributions to Music Education43, 137–158 (2018),https: //www.jstor.org/stable/26478003

    Hanson, J.: Assessing the educational value of youtube videos for beginning in- strumental music. Contributions to Music Education43, 137–158 (2018),https: //www.jstor.org/stable/26478003

    2018

  11. [20]

    In: Proc

    Hu, S., Zhang, B., Lu, J., Jiang, Y., Wang, W., Kong, L., Zhao, W., Jiang, T.: WideResNet with Joint Representation Learning and Data Augmentation for Cover Song Identification. In: Proc. Interspeech 2022. pp. 4187–4191 (2022). https://doi.org/10.21437/Interspeech.2022-10600 O...

    2022 doi

  12. [21]

    In: International Society for Music Information Retrieval Conference (ISMIR) (2007)

    Jones, M.C., Downie, J.S., Ehmann, A.F.: Human similarity judgments: Impli- cations for the design of formal evaluations. In: International Society for Music Information Retrieval Conference (ISMIR) (2007)

    2007

  13. [22]

    In: International Society for Music Information Retrieval Conference (ISMIR) (2019)

    Li, B., Kumar, A.: Query by video: Cross-modal music retrieval. In: International Society for Music Information Retrieval Conference (ISMIR) (2019)

    2019

  14. [23]

    arXiv preprint arXiv:2004.00584 (2020)

    Li, Y., Li, J., Suhara, Y., Doan, A., Tan, W.C.: Deep entity matching with pre- trained language models. arXiv preprint arXiv:2004.00584 (2020)

    2020 arXiv

  15. [24]

    Computers in Human Behavior 50, 108–124 (2015).https://doi.org/10.1016/j.chb.2015.01.067

    Liikkanen, L.A., Salovaara, A.: Music on youtube: User engagement with tradi- tional, user-appropriated and derivative videos. Computers in Human Behavior 50, 108–124 (2015).https://doi.org/10.1016/j.chb.2015.01.067

    2015 doi

  16. [25]

    In: 2023 IEEE International Conference on Mul- timedia and Expo (ICME)

    Liu, F., Tuo, D., Xu, Y., Han, X.: Coverhunter: Cover song identification with refined attention and alignments. In: 2023 IEEE International Conference on Mul- timedia and Expo (ICME). pp. 1080–1085. IEEE (2023)

    2023

  17. [26]

    YouTube and Music22(4), 169 (2016)

    Martet, S.: The circulation of user-appropriated music content on youtube. YouTube and Music22(4), 169 (2016)

    2016

  18. [27]

    European Journal of Marketing 53 (07 2018)

    Matherly, T.: A panel for lemons? positivity bias, reputation systems and data quality on mturk. European Journal of Marketing 53 (07 2018). https://doi. org/10.1108/EJM-07-2017-0491

    2018 doi

  19. [28]

    New Media & Society23(6), 1624–1641 (2021)

    McDaniel, B.: Popular music reaction videos: Reactivity, creator labor, and the performance of listening online. New Media & Society23(6), 1624–1641 (2021). https://doi.org/10.1177/1461444820918549

    2021 doi

  20. [29]

    Addiction115 (03 2020).https://doi.org/10.1111/ add.15032

    Mellis,A.,Bickel,W.:Mechanicalturkdatacollectioninaddictionresearch:Utility, concerns and best practices. Addiction115 (03 2020).https://doi.org/10.1111/ add.15032

    2020

  21. [30]

    arXiv preprint arXiv:1803.10963 (2018)

    Okabe,K.,Koshinaka,T.,Shinoda,K.:Attentivestatisticspoolingfordeepspeaker embedding. arXiv preprint arXiv:1803.10963 (2018)

    2018 arXiv

  22. [31]

    Behavior Research Methods46, 1023–1031 (12 2013)

    Peer, E., Vosgerau, J., Acquisti, A.: Reputation as a sufficient condition for data quality on Amazon Mechanical Turk. Behavior Research Methods46, 1023–1031 (12 2013). https://doi.org/10.3758/s13428-013-0434-y

    2013 doi

  23. [32]

    In: ISMIR

    Silva, D.F., de Souza, V.M., Batista, G.E.: Music shapelets for fast cover song recognition. In: ISMIR. pp. 441–447 (2015)

    2015

  24. [33]

    In: International Conference on Multimedia and Expo (ICME)

    Smith, J.B.L., Hamasaki, M., Goto, M.: Classifying derivative works with search, text, audio and video features. In: International Conference on Multimedia and Expo (ICME). pp. 1422–1427. IEEE (2017). https://doi.org/10.1109/ICME. 2017.8019444

    2017

  25. [34]

    In: 2018 IEEE International Conference on Multi- media and Expo (ICME)

    Xu, X., Chen, X., Yang, D.: Key-invariant convolutional neural network toward ef- ficient cover song identification. In: 2018 IEEE International Conference on Multi- media and Expo (ICME). pp. 1–6 (2018).https://doi.org/10.1109/ICME.2018. 8486531

    2018 doi

  26. [35]

    IEEE Signal Processing Magazine 38(6), 115–136 (2021).https://doi.org/10.1109/MSP.2021.3105941

    Yesiler, F., Doras, G., Bittner, R.M., Tralie, C.J., Serrà, J.: Audio-based musical version identification: Elements and challenges. IEEE Signal Processing Magazine 38(6), 115–136 (2021).https://doi.org/10.1109/MSP.2021.3105941

    2021

  27. [36]

    In: International Conference on Acous- tics, Speech and Signal Processing (ICASSP)

    Yesiler, F., Serrà, J., Gómez, E.: Accurate and scalable version identification using musically-motivated embeddings. In: International Conference on Acous- tics, Speech and Signal Processing (ICASSP). pp. 21–25. IEEE (2020). https: //doi.org/10.1109/ICASSP40776.2020.9053793

    2020

  28. [37]

    In: Proc

    Yesiler, F., Serrà, J., Gómez, E.: Less is more: Faster and better music version identification with embedding distillation. In: Proc. of the Int. Soc. for Music In- formation Retrieval Conf. (ISMIR) (2020) 22 Simon Hachmeier and Robert Jäschke

    2020

  29. [38]

    In: Proc

    Yesiler, F., Tralie, C., Correya, A., Silva, D.F., Tovstogan, P., Gómez, E., Serra, X.: Da-TACOS: A dataset for cover song identification and understanding. In: Proc. of the 20th Int. Soc. for Music Information Retrieval Conf. (ISMIR). pp. 327–334. Delft, The Netherlands (2019)

    2019

  30. [39]

    In: ICASSP 2020-2020 IEEE Inter- national Conference on Acoustics, Speech and Signal Processing (ICASSP)

    Yu, Z., Xu, X., Chen, X., Yang, D.: Learning a representation for cover song iden- tification using convolutional neural network. In: ICASSP 2020-2020 IEEE Inter- national Conference on Acoustics, Speech and Signal Processing (ICASSP). pp. 541–545. IEEE (2020)

    2020

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.