Pith. sign in

REVIEW 4 major objections 6 minor 37 references

MQAD: A Large-Scale Question Answering Dataset for Training Music Large Language Models

T0 review · 4 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read MQAD supplies nearly three million music QA pairs and shows they improve music captioning and question answering.

desk verdict A big, genuinely new music QA dataset whose usefulness is undercut by circular in-domain evaluation and the absence of any human audit of automatically generated answers. read the letter →

arxiv 2508.19514 v1 pith:TSPMC4KC submitted 2025-08-27 cs.SD

Zhihao Ouyang , Ju-Chiang Wang , Daiyu Zhang , Bin Chen , Shangjie Li , Quan Lin This is my paper · ORCID
classification cs.SD
keywords musicquestionansweringcaptioningmultimodallargelanguagemodelMillionSongDatasetInformationRetrievaltemporalstructureGPT-4evaluationQA
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 claims that a new dataset, MQAD, built on the Million Song Dataset, gives music-language models a resource they have lacked: nearly three million question-answer pairs and captions that include time-varying musical details such as chords, sections, beats, and keys. The authors argue that this scale and temporal structure let a multimodal model trained on MQAD beat prior captioning approaches on standard music captioning tests, with the best fine-tuned model reaching BLEU-1 of 30.30 against 29.09 for the strongest previous method. If true, MQAD is a training resource for music LLMs to answer structural and compositional questions rather than only tag mood or genre.

What carries the argument

The engine is a two-stage data-generation pipeline. In stage one, transformer-based Music Information Retrieval (MIR) models extract beats, chords, keys, structural sections, and multi-instrument transcriptions from each audio clip, and these events are serialized into text such as timestamped chord names and section labels. In stage two, GPT-4 Turbo, prompted with expert-authored meta questions and a retrieval-augmented generation module, converts that text into diverse QA pairs and captions. On the model side, MMQAD couples Whisper as a frozen audio encoder with LLaMA2-7B as the language backbone, trained with LoRA to keep compute costs down.

What would settle it

Take 500 MQAD QA pairs sampled across chord, key, structure, and transcription questions and have trained musicians check the stated facts against the audio; if the fraction of verifiably wrong answers approaches the benchmark error rates of the MIR models for chords and structure, then models trained on MQAD are being taught machine errors rather than musical facts.

Watch

Extended reading notes

Core claim

The paper claims that a fully automatic pipeline can turn audio into a large, usable QA training corpus, and that training on that corpus transfers to better music captioning and question answering. Concretely, MQAD contains 4 question-answer pairs and 7 captions per track across 270,000 tracks, totaling about 3 million items; the released QA subset has about 804,000 training QA pairs. The authors report that fine-tuning on MQAD improves supervised music captioning (MMQAD-C+F: BLEU-1 30.30 vs. 29.09 for LP-MusicCaps) and that MQAD-trained models score far higher on MQAD's own QA test set (BLEU-1 51.86 vs. 15.24 for LP-MusicCaps). They also introduce a GPT-4 Turbo judge that rates model outputs across eight musical dimensions and report high scores for their models on that pseudo-subjective metric.

Load-bearing premise

The whole dataset rests on treating outputs of automatic MIR models as correct facts and on GPT-4-generated QA text as a faithful rendering of those facts, with no human verification of either step in the construction pipeline.

Editorial extensions

If this is right

  • If MQAD's QA pairs are reliable, the music-AI community gets a way to train models that answer questions about chords, sections, and beat structure, not just genre and mood.
  • A model pre-trained on MQAD alone outperforms one pre-trained on the larger but caption-only LP-MusicCaps-MSD corpus, suggesting that question diversity can matter more than raw caption count for music-language pretraining.
  • Since MQAD-Full includes 7 captions per track beyond the released QA subset, the remaining 64% of the dataset is a ready resource for scaling captioning and text-to-music work.
  • The GPT-4-based subjective metric offers a cheap, repeatable proxy for evaluating music QA along multiple dimensions, which could become a standard evaluation tool.

Reading between the lines

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

  • A human audit of a random sample of MQAD QA pairs would be the natural next test: the dataset's truthfulness as ground truth depends on the unverified correctness of the MIR extractors, and chord and structure labels are likely noisier than genre tags.
  • The large verbosity gap in MMQAD outputs (about 75 tokens versus 50 for LP-MusicCaps) may inflate semantic similarity scores while deflating long-gram BLEU scores, so part of the reported gain may be a style effect rather than purely factual understanding.
  • If the pipeline proves reliable, the same recipe of serializing MIR events and asking an LLM to generate QA pairs could transfer to other audio domains where temporal structure matters, such as sound events or speech prosody.
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

4 major / 6 minor

Summary. The paper introduces MQAD, a large-scale music question-answering and captioning dataset built on the Million Song Dataset (MSD). The dataset is constructed by running automatic MIR models (beat tracking, chord/key detection, structure segmentation, and multi-instrument transcription) on roughly 270,000 tracks and then prompting GPT-4 Turbo with the extracted features to generate nearly 3 million QA pairs and captions. The authors also present MMQAD, a multimodal LLM that combines LLaMA2-7B and Whisper, trained on MQAD. They report that MMQAD achieves state-of-the-art music captioning (e.g., BLEU-1 30.30 vs. 29.09 for LP-MusicCaps on the LP-MusicCaps-MC test set) and strong performance on an in-house QA test set. The dataset and code are publicly released.

Significance. If the dataset's quality and the evaluation protocol were reliable, MQAD would be a valuable community resource: it is an order of magnitude larger than existing music captioning datasets and uniquely includes time-varying structural and harmonic information. The open release of the dataset and model is a strength. However, the current manuscript does not establish that the generated QA pairs are factually correct, and the main evaluation is conducted on a test set generated by the same pipeline used to create the training data. The reported captioning gains are small, mixed across metrics, and lack statistical characterization. The paper's central claims therefore rest on assumptions that need to be tested and reported.

major comments (4)
  1. [II.A (Feature Extraction and QA Pair Generation)] The dataset is built entirely from automatic MIR predictions (beat, chord, key, structure, transcription) that are then converted into QA text by GPT-4 Turbo, yet no human verification or error analysis is reported for either stage. Chord recognition, key detection, structural segmentation, and polyphonic transcription are known to be imperfect, and LLM conversion can introduce additional hallucination. Because the entire value of MQAD depends on the factual correctness of the answers, the paper must either report a human-annotation study (e.g., agreement rates on a sampled subset for each MIR category and question type) or clearly state and quantify this limitation. Without such evidence, the claim that MQAD is a high-quality factual music QA resource is not supported.
  2. [III.C and Tables III–IV] The MQAD test set (110K QA pairs) is generated by the same MIR-plus-LLM pipeline used to create the training data, so the large score gaps between MMQAD models and LP-MusicCaps on this test set largely reflect in-distribution fit to the generator's question distribution and MIR predictions, not musical understanding. The subjective evaluation in Table IV uses GPT-4 Turbo as the judge, which is the same model family that generated the QA pairs, compounding the circularity. Additionally, the description of this subjective metric (Section II.C) does not provide the prompt template or scoring rubric, making it impossible to reproduce. The authors should evaluate on a human-verified test set or at least on a subset whose ground truth has been manually checked, and they should provide full details of the LLM-as-judge protocol.
  3. [Table II] The headline captioning improvement is marginal: BLEU-1 increases from 29.09 (LP-MusicCaps) to 30.30 (MMQAD-C+F), while METEOR actually decreases (22.39 to 22.25), and no confidence intervals, multiple-seed variance, or statistical significance tests are reported. Given the paper's claim of state-of-the-art performance, the authors should provide error bars or significance tests across random seeds, and discuss the metric trade-offs rather than selectively highlighting BLEU-1.
  4. [Abstract and Section III.B] The claim of "state-of-the-art" captioning is not supported by the comparison set, which includes only LP-MusicCaps and a few older baselines (Tag Concat, Template, K2C-Aug). No comparison is made to other recent music MLLMs such as MU-LLaMA or LLark under the same transfer/fine-tuning protocol. At minimum, the manuscript should restrict its claim to "improvements over the compared baselines" or add the missing comparisons.
minor comments (6)
  1. [Abstract] The sentence "The dataset and codes are at" should be rephrased (e.g., "The dataset and code are available at").
  2. [II.A] The description of "Meta Questions from Music Experts" does not specify who the experts are or how many questions were drafted; this information is needed to assess the diversity and coverage of the generated QA pairs.
  3. [II.C] The paper calls the GPT-4-based evaluation metric "novel," but LLM-as-judge evaluation is a standard practice in NLP; the novelty claim should be toned down, and the metric should be validated against human judgments, which is currently missing.
  4. [Table I] The column heading "C/A" is ambiguous; clarify that it denotes the number of captions/answers per audio clip. Also, the text says MQAD-Full comprises 4 QA pairs and 7 captions per track, but the table does not break these out separately; a clearer breakdown would help.
  5. [II.B] The model name is inconsistently written as "LLAMA2" and "LLaMA 2" in different places; standardize the spelling.
  6. [III.A] When defining MMQAD-B, the description "combination of LP-MusicCaps-MSD training set and MQAD training set" should state whether the MQAD split used is the QA subset or the full dataset, since the subsequent analysis (e.g., Table III) depends on this detail.

Circularity Check

2 steps flagged · score 6.0 of 10

MQAD's in-house QA evaluation is partially circular: the test labels and the GPT-4 judge come from the same automatic generation pipeline that created the training set, while the external captioning benchmark provides only modest independent grounding.

  1. fitted input called prediction [Section II.A (QA Pair Generation) and Section III.C (Results for the Music QA Task)]
    "We used an LLM to generate text QA pairs from the extracted musical event data and the meta-information provided by MSD. ... we compiled the MQAD test set, which includes 100K samples featuring detailed MIR questions such as chord progression and music structure."

    The MQAD training QA pairs and the MQAD test QA pairs are both produced by the same pipeline: automatic MIR feature extraction followed by GPT-4 Turbo generation with the same RAG/backbone prompts. A model trained on this distribution can score highly on the test set by reproducing the generator's question style and its MIR-derived 'facts'; the test set contains no independent human-verified ground truth. Thus Table III's QA scores measure fit to the generator distribution, not verified musical understanding.

  2. other [Section II.C (Subjective and Objective Metrics)]
    "Specifically, we employed GPT4-Turbo to compare predicted answers with the ground truth across eight distinct musical dimensions"

    The subjective evaluator is the same model family (GPT-4 Turbo) that generated the MQAD QA pairs in Section II.A. Since the 'ground truth' answers are themselves GPT-4-generated text, the Table IV scores largely measure whether model outputs are stylistically consistent with GPT-4's own writing rather than independent musical correctness. This self-referential loop weakens the in-house QA evaluation, although it does not affect the external LP-MusicCaps-MC captioning comparison.

full rationale

The paper's core dataset is a large synthetic QA corpus: Section II.A generates questions/captions from MIR model outputs (beat, chord, key, structure, transcription) using GPT-4 Turbo. The central circularity is in evaluation: the MQAD test and validation sets are drawn from the same generation pipeline, and the subjective judge is the same model family that generated the data. Consequently, Tables III and IV primarily demonstrate that models trained on MQAD can match the MQAD generator's distribution and style, not that the model has learned musically verified facts. This is a partial-by-construction reduction of the QA 'prediction' claim. In contrast, the LP-MusicCaps-MC captioning benchmark is external and independent, but the reported advantage is small (BLEU-1 30.30 vs 29.09), METEOR is lower for the fine-tuned model, and no statistical significance is reported; it therefore only weakly supports the broader claim of advancement. The MIR feature-extraction citations are peer-reviewed external papers and are not load-bearing self-citations in a circular sense; the issue is the absence of human verification and the self-generated evaluation, not the feature extractors themselves. Overall, the derivation of the dataset is non-circular, but the in-house QA validation is substantially circular, giving a score of 6.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

The dataset's validity rests on unverified automatic feature extraction and unvalidated LLM generation; these are domain assumptions rather than free parameters. The evaluation also assumes GPT-4 can stand in for human raters.

assumptions (5)
  • domain assumption MIR model predictions for beat, chord, key, structure, and transcription are accurate enough to serve as ground truth.
    Section II.A uses these predictions as the only factual layer for QA generation; no human validation or error analysis is reported.
  • domain assumption GPT-4 Turbo generated QA pairs are factually consistent with the MIR text and span diverse questions.
    Section II.A relies on prompts, RAG, and temperature 0.95 to ensure diversity and correctness, but no manual audit or inter-annotator agreement is given.
  • domain assumption The track selection criterion from [24] yields roughly 20% of MSD without introducing bias.
    Section II.A Data source states selection based on [24] but does not describe the criterion; potential selection bias is unexamined.
  • domain assumption Appending 'write a music caption for this track' makes LP-MusicCaps-MC test a fair QA benchmark for MMQAD.
    Section III.A adapts the test set this way while the LP-MusicCaps baseline receives only audio; the prompt asymmetry is not discussed.
  • domain assumption GPT-4 Turbo subjective judgments correlate with human preferences for music QA.
    Section II.C cites NLP findings from [18]-[21] but conducts no human study on music outputs.

how reviews work

0 comments
Cite this review

Pith. "Pith review of MQAD: A Large-Scale Question Answering Dataset for Training Music Large Language Models." pith.science (2026). https://pith.science/paper/TSPMC4KC

@misc{pith2026250819514,
  author       = {Pith},
  title        = {Pith review of: MQAD: A Large-Scale Question Answering Dataset for Training Music Large Language Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TSPMC4KC}},
  note         = {Machine review of arXiv:2508.19514}
}
read the original abstract

Question-answering (QA) is a natural approach for humans to understand a piece of music audio. However, for machines, accessing a large-scale dataset covering diverse aspects of music is crucial, yet challenging, due to the scarcity of publicly available music data of this type. This paper introduces MQAD, a music QA dataset built on the Million Song Dataset (MSD), encompassing a rich array of musical features, including beat, chord, key, structure, instrument, and genre -- across 270,000 tracks, featuring nearly 3 million diverse questions and captions. MQAD distinguishes itself by offering detailed time-varying musical information such as chords and sections, enabling exploration into the inherent structure of music within a song. To compile MQAD, our methodology leverages specialized Music Information Retrieval (MIR) models to extract higher-level musical features and Large Language Models (LLMs) to generate natural language QA pairs. Then, we leverage a multimodal LLM that integrates the LLaMA2 and Whisper architectures, along with novel subjective metrics to assess the performance of MQAD. In experiments, our model trained on MQAD demonstrates advancements over conventional music audio captioning approaches. The dataset and code are available at https://github.com/oyzh888/MQAD.

Figures

Figures reproduced from arXiv: 2508.19514 by the authors.

Figure 1
Figure 1. The MQAD dataset offers diverse coverage of fine-grained MIR facets, [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. MQAD dataset construction and MMQAD model training. [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

37 extracted references · 15 canonical work pages

  1. [1]

    Qwen-vl: A frontier large vision-language model with versatile abilities,

    J. Bai, S. Bai, S. Yang, S. Wang, S. Tan, P. Wang, J. Lin, C. Zhou, and J. Zhou, “Qwen-vl: A frontier large vision-language model with versatile abilities,” arXiv preprint arXiv:2308.12966 , 2023

    arXiv 2023

  2. [2]

    Qwen-audio: Advancing universal audio understanding via unified large-scale audio-language models,

    Y . Chu, J. Xu, X. Zhou, Q. Yang, S. Zhang, Z. Yan, C. Zhou, and J. Zhou, “Qwen-audio: Advancing universal audio understanding via unified large-scale audio-language models,” arXiv preprint arXiv:2311.07919 , 2023

    arXiv 2023

  3. [3]

    Visual instruction tuning,

    H. Liu, C. Li, Q. Wu, and Y . J. Lee, “Visual instruction tuning,”NeurIPS, vol. 36, 2024

    2024

  4. [4]

    Music understanding llama: Advancing text-to-music generation with question answering and captioning,

    S. Liu, A. S. Hussain, C. Sun, and Y . Shan, “Music understanding llama: Advancing text-to-music generation with question answering and captioning,” in ICASSP, 2024, pp. 286–290

    work page 2024

  5. [5]

    The dawn of lmms: Preliminary explorations with gpt-4v (ision),

    Z. Yang, L. Li, K. Lin, J. Wang, C.-C. Lin, Z. Liu, and L. Wang, “The dawn of lmms: Preliminary explorations with gpt-4v (ision),” arXiv preprint arXiv:2309.17421, 2023

    arXiv 2023

  6. [6]

    Sparks of artificial general intelligence: Early experiments with gpt-4,

    S. Bubeck, V . Chandrasekaran, R. Eldan, J. Gehrke, E. Horvitz, E. Kamar, P. Lee, Y . T. Lee, Y . Li, S. Lundberg, et al., “Sparks of artificial general intelligence: Early experiments with gpt-4,” arXiv preprint arXiv:2303.12712, 2023

    arXiv 2023

  7. [7]

    Emergent abilities of large language models,

    J. Wei, Y . Tay, R. Bommasani, C. Raffel, B. Zoph, S. Borgeaud, D. Yogatama, M. Bosma, D. Zhou, D. Metzler, et al., “Emergent abilities of large language models,” arXiv preprint arXiv:2206.07682 , 2022

    arXiv 2022

  8. [8]

    Are emergent abilities of large language models a mirage?,

    R. Schaeffer, B. Miranda, and S. Koyejo, “Are emergent abilities of large language models a mirage?,” Advances in Neural Information Processing Systems, vol. 36, 2024

    work page 2024

Show all 37 references
  1. [9]

    MiniGPT-4: Enhancing vision-language understanding with advanced large language models,

    D. Zhu, J. Chen, X. Shen, X. Li, and M. Elhoseiny, “MiniGPT-4: Enhancing vision-language understanding with advanced large language models,” arXiv preprint arXiv:2304.10592 , 2023

    2023 arXiv

  2. [10]

    Gemini: a family of highly capable multimodal models,

    G. Team, R. Anil, S. Borgeaud, Y . Wu, J.-B. Alayrac, J. Yu, R. Soricut, J. Schalkwyk, A. M. Dai, A. Hauth, et al., “Gemini: a family of highly capable multimodal models,” arXiv preprint arXiv:2312.11805 , 2023

    2023 arXiv

  3. [11]

    GPT-4 technical report,

    J. Achiam, S. Adler, S. Agarwal, L. Ahmad, I. Akkaya, F. L. Aleman, D. Almeida, J. Altenschmidt, S. Altman, S. Anadkat, et al., “GPT-4 technical report,” arXiv preprint arXiv:2303.08774 , 2023

    2023 arXiv

  4. [12]

    Robust speech recognition via large-scale weak supervision,

    A. Radford, J. W. Kim, T. Xu, G. Brockman, C. McLeavey, and I. Sutskever, “Robust speech recognition via large-scale weak supervision,” in International Conference on Machine Learning . PMLR, 2023, pp. 28492–28518

    2023

  5. [13]

    Salmonn: Towards generic hearing abilities for large language models,

    C. Tang, W. Yu, G. Sun, X. Chen, T. Tan, W. Li, L. Lu, Z. Ma, and C. Zhang, “Salmonn: Towards generic hearing abilities for large language models,” arXiv preprint arXiv:2310.13289 , 2023

    2023 arXiv

  6. [14]

    LLark: A multimodal foundation model for music,

    J. Gardner, S. Durand, D. Stoller, and R. Bittner, “LLark: A multimodal foundation model for music,” arXiv preprint arXiv:2310.07160 , 2023

    2023 arXiv

  7. [15]

    LP-MusicCaps: LLM-based pseudo music captioning,

    S. Doh, K. Choi, J. Lee, and J. Nam, “LP-MusicCaps: LLM-based pseudo music captioning,” in ISMIR, 2023

    2023

  8. [16]

    The million song dataset,

    T. Bertin-Mahieux, D. P. Ellis, B. Whitman, and P. Lamere, “The million song dataset,” in ISMIR, 2011

    2011

  9. [17]

    LLaMA 2: Open foundation and fine-tuned chat models,

    H. Touvron, L. Martin, K. Stone, P. Albert, A. Almahairi, Y . Babaei, N. Bashlykov, S. Batra, P. Bhargava, S. Bhosale, et al., “LLaMA 2: Open foundation and fine-tuned chat models,” arXiv preprint arXiv:2307.09288, 2023

    2023 arXiv

  10. [18]

    Gpteval: Nlg evaluation using gpt-4 with better human alignment,

    Y . Liu, D. Iter, Y . Xu, S. Wang, R. Xu, and C. Zhu, “Gpteval: Nlg evaluation using gpt-4 with better human alignment,” arXiv preprint arXiv:2303.16634, 2023

    2023 arXiv

  11. [19]

    GPT-4 as an effective zero-shot evaluator for scientific figure captions,

    T.-Y . Hsu, C.-Y . Huang, R. Rossi, S. Kim, C. L. Giles, and T.-H. K. Huang, “GPT-4 as an effective zero-shot evaluator for scientific figure captions,” arXiv preprint arXiv:2310.15405 , 2023

    2023 arXiv

  12. [20]

    Is GPT-4 a reliable rater? evaluating consistency in gpt-4 text ratings,

    V . Hackl, A. E. M ¨uller, M. Granitzer, and M. Sailer, “Is GPT-4 a reliable rater? evaluating consistency in gpt-4 text ratings,” arXiv preprint arXiv:2308.02575, 2023

    2023 arXiv

  13. [21]

    Large language models are human-level prompt engineers,

    Y . Zhou, A. I. Muresanu, Z. Han, K. Paster, S. Pitis, H. Chan, and J. Ba, “Large language models are human-level prompt engineers,” arXiv preprint arXiv:2211.01910, 2022

    2022 arXiv

  14. [22]

    Learning to answer questions in dynamic audio-visual scenarios,

    G. Li, Y . Wei, Y . Tian, C. Xu, J.-R. Wen, and D. Hu, “Learning to answer questions in dynamic audio-visual scenarios,” in CVPR, 2022, pp. 19108–19118

    2022

  15. [23]

    Musicagent: An ai agent for music understanding and generation with large language models,

    D. Yu, K. Song, P. Lu, T. He, X. Tan, W. Ye, S. Zhang, and J. Bian, “Musicagent: An ai agent for music understanding and generation with large language models,” arXiv preprint arXiv:2310.11954 , 2023

    2023 arXiv

  16. [24]

    Multi-level and multi-scale feature aggregation using pretrained convolutional neural networks for music auto-tagging,

    J. Lee and J. Nam, “Multi-level and multi-scale feature aggregation using pretrained convolutional neural networks for music auto-tagging,” IEEE signal processing letters , vol. 24, no. 8, pp. 1208–1212, 2017

    2017

  17. [25]

    Modeling beats and downbeats with a time-frequency transformer,

    Y .-N. Hung, J.-C. Wang, X. Song, W.-T. Lu, and M. Won, “Modeling beats and downbeats with a time-frequency transformer,” in ICASSP, 2022, pp. 401–405

    2022

  18. [26]

    SpecTNT: A time-frequency transformer for music audio,

    W.-T. Lu, J.-C. Wang, M. Won, K. Choi, and X. Song, “SpecTNT: A time-frequency transformer for music audio,” in ISMIR, 2021

    2021

  19. [27]

    To catch a chorus, verse, intro, or anything else: Analyzing a song with structural functions,

    J.-C. Wang, Y .-N. Hung, and J. B. Smith, “To catch a chorus, verse, intro, or anything else: Analyzing a song with structural functions,” in ICASSP, 2022, pp. 416–420

    2022

  20. [28]

    Multitrack music transcription with a time-frequency perceiver,

    W.-T. Lu, J.-C. Wang, and Y .-N. Hung, “Multitrack music transcription with a time-frequency perceiver,” in ICASSP, 2023

    2023

  21. [29]

    M. F. Matthew E. P. Davies, Sebastian Bock, Tempo, Beat and Downbeat Estimation, ISMIR Tutorial, Nov. 2021

    2021

  22. [30]

    20 years of automatic chord recognition from audio,

    J. Pauwels, K. O’Hanlon, E. G ´omez, M. Sandler, et al., “20 years of automatic chord recognition from audio,” in ISMIR, 2019

    2019

  23. [31]

    MuSFA: Improving music structural function analysis with partially labeled data,

    J.-C. Wang, J. B. Smith, and Y .-N. Hung, “MuSFA: Improving music structural function analysis with partially labeled data,” ISMIR Late Breaking & Demo , 2022

    2022

  24. [32]

    Retrieval-augmented generation for large language models: A survey,

    Y . Gao, Y . Xiong, X. Gao, K. Jia, J. Pan, Y . Bi, Y . Dai, J. Sun, and H. Wang, “Retrieval-augmented generation for large language models: A survey,” arXiv preprint arXiv:2312.10997 , 2023

    2023 arXiv

  25. [33]

    AudioCaps: Generating captions for audios in the wild,

    C. D. Kim, B. Kim, H. Lee, and G. Kim, “AudioCaps: Generating captions for audios in the wild,” in Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, 2019

    2019

  26. [34]

    Large-scale contrastive language-audio pretraining with feature fusion and keyword-to-caption augmentation,

    Y . Wu, K. Chen, T. Zhang, Y . Hui, T. Berg-Kirkpatrick, and S. Dubnov, “Large-scale contrastive language-audio pretraining with feature fusion and keyword-to-caption augmentation,” in ICASSP, 2023

    2023

  27. [35]

    WavCaps: A ChatGPT-assisted weakly-labelled audio captioning dataset for audio-language multimodal research,

    X. Mei, C. Meng, H. Liu, Q. Kong, T. Ko, C. Zhao, M. D. Plumbley, Y . Zou, and W. Wang, “WavCaps: A ChatGPT-assisted weakly-labelled audio captioning dataset for audio-language multimodal research,” arXiv preprint arXiv:2303.17395, 2023

    2023 arXiv

  28. [36]

    Musiclm: Generating music from text,

    A. Agostinelli, T. I. Denk, Z. Borsos, J. Engel, M. Verzetti, A. Caillon, Q. Huang, A. Jansen, A. Roberts, M. Tagliasacchi, et al., “Musiclm: Generating music from text,” arXiv preprint arXiv:2301.11325 , 2023

    2023 arXiv

  29. [37]

    LoRa: Low-rank adaptation of large language models,

    E. J. Hu, Y . Shen, P. Wallis, Z. Allen-Zhu, Y . Li, S. Wang, L. Wang, and W. Chen, “LoRa: Low-rank adaptation of large language models,” arXiv preprint arXiv:2106.09685 , 2021

    2021 arXiv

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.