Pith. sign in

REVIEW 4 major objections 5 minor 34 references

MIMII-Agent: Leveraging LLMs with Function Calling for Relative Evaluation of Anomalous Sound Detection

T0 review · 4 major / 5 minor · reviewed 2026-08-06 · 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 LLM-selected audio transformations applied to normal machine sounds produce synthetic anomalies whose per-machine detection-difficulty ranking matches real anomalies, so anomalous sound detection systems can be…

desk verdict A practical LLM-based pipeline for synthesizing anomalies with a useful ablation, but the validation is too thin and an untested domain-shift confound weakens the central claim. read the letter →

arxiv 2507.20666 v1 pith:SCMT27FB submitted 2025-07-28 eess.AS cs.AIcs.LGcs.SD

Harsh Purohit , Tomoya Nishida , Kota Dohi , Takashi Endo , Yohei Kawaguchi This is my paper · ORCID
classification eess.AScs.AIcs.LGcs.SD
keywords anomaloussounddetectionunsupervisedanomalylargelanguagemodelsfunctioncallingsyntheticgenerationrelativeevaluationaudiotransformationsmachineconditionmonitoring
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

Unsupervised anomalous sound detection systems are normally judged by how well they separate normal from faulty sounds, but real fault recordings are rare and uneven in difficulty. This paper proposes judging such systems relatively, by ranking their detection performance across machine types, and argues that this ranking can be obtained from synthetic faults alone. The synthesis method has a large language model read a caption describing a machine's operating condition and choose an audio-transformation function—adding squeaking, rattling, humming, or similar effects—to apply to normal machine audio. With five machine types and a detector trained only on normal sounds, the ranking of detection difficulty from these synthetic anomalies matched the ranking from real anomalies under two scoring methods. If the finding holds, maintenance engineers can learn where a monitoring system is trustworthy before any real faults have been recorded.

What carries the argument

The load-bearing mechanism is the LLM function-calling loop. A prompt holding the machine caption and a library of hand-coded audio transformations—add_squeaking, add_rattling, add_grinding, add_humming, and similar—is sent to the model, which returns the name of the transformation most appropriate to the described fault; that function is then applied to normal audio to produce an anomalous clip. The evaluative side is the rank comparison: per-machine AUC scores from synthetic anomalies are ordered and compared with per-machine AUC scores from real anomalies, so the quantity that matters is the ordering rather than the score value. Supporting the pipeline are a caption generator that turns metadata into text and a generative model that supplies the normal audio clips.

What would settle it

Resample the real anomaly sets many times, recompute the per-machine AUC ranking for each resample, and check whether the synthetic ranking falls inside the resulting distribution; if the match disappears, or if the ranking changes when pumps and motors are added to the five machine types, the central claim would be falsified. A cheaper check is to compute rank correlation between synthetic and real AUCs across many more real anomalies per machine type and see whether it stays near perfect with uncertainty bounds.

Watch

Extended reading notes

Core claim

The central claim is that relative detection difficulty across machine types is preserved when anomalies are synthesized by LLM-guided audio transformations instead of recorded from real faults. The paper defines relative evaluation as the ordering of per-machine AUC (area under the ROC curve) scores, arguing that absolute AUC depends on the severity of the test anomalies while the ordering is more stable. In the experiment, a single autoencoder trained only on normal sounds is scored on five machine types—bearing, gearbox, fan, slide rail, valve—using both synthetic and real anomalies. The observed rank order is the same in both cases under both mean squared error and Mahalanobis distance scoring: fan is easiest, followed by gearbox, bearing, slide rail, and valve hardest. The ablation study adds that random selection of audio effects destroys the rank match while manual keyword-based selection preserves it, which the paper takes as evidence that the LLM's contextual understanding, rather than the transformations by themselves, is what makes the synthetic anomalies informative.

Load-bearing premise

The load-bearing premise is that a ranking computed from 50 synthetic and 50 real anomalies per machine type, evaluated with a single autoencoder and no uncertainty estimates, correctly captures how difficult each machine type is to monitor.

Editorial extensions

If this is right

  • A UASD system's relative strengths and weaknesses across machine types can be assessed without collecting real fault sounds for every machine type.
  • Benchmarking can move from absolute AUC to rank-based comparison, which is less sensitive to the severity of the anomalies used in the test set.
  • Adding a new machine type to an evaluation suite only requires describing its typical faults in text and having normal operating audio available.
  • Contextual understanding is doing real work in the pipeline: random anomaly selection produced rankings uncorrelated with real anomalies, while LLM and manual keyword selection agreed with them.
  • Because the same rank order appeared under MSE and Mahalanobis scoring, the relative-evaluation signal does not depend on the choice of anomaly score.

Reading between the lines

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

  • If the result generalizes beyond these five machine types, rank-based benchmarks could replace absolute AUC in unsupervised anomalous sound detection, letting practitioners compare systems without large labeled fault corpora.
  • The synthesis is limited by what can be expressed in text and by the hand-coded transformation library, so failure modes with no obvious linguistic description or audio signature may be under-represented.
  • A natural next experiment would vary the LLM, the prompt template, and the transformation library to test whether rank preservation is a stable property of the pipeline or an accident of the specific choices.
  • The same text-to-transformation pattern could plausibly transfer to other condition-monitoring signals, such as vibration, current, or thermal data, where fault descriptions select transformations rather than raw samples.
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 / 5 minor

Summary. The paper proposes MIMII-Agent, a pipeline that uses an LLM with function calling to map textual machine descriptions to audio DSP effects applied to MIMII-Gen-generated normal sounds, creating synthetic anomalies for 'relative evaluation' of UASD systems (rankings of AUC across machine types) without collecting real fault data. The method is tested with one autoencoder trained on real normal sounds, evaluating five machine types (bearing, gearbox, fan, slide rail, valve) using both synthetic and real anomalies, and an ablation comparing GPT-4o selection to manual keyword mapping and random selection. The paper reports consistent AUC rankings between synthetic and real anomalies and concludes that LLM-based synthesis can reliably benchmark relative detection difficulty.

Significance. If the result is robust, the paper offers a scalable, low-cost alternative to collecting real fault sounds for system benchmarking, which would be of practical value for industrial condition monitoring and DCASE-style evaluations. The paper is clearly structured and includes a useful ablation with a random baseline; however, the current evidence does not yet rule out a domain-shift confound from using MIMII-Gen normal audio as the synthesis base, and the statistical support (one autoencoder run, n=5) is thin. The cross-system claim is also overstated.

major comments (4)
  1. [§4.3, Table 3] The synthetic-anomaly AUC is measured by scoring MIMII-Gen normal audio with added effects against real normal audio, while the autoencoder in §4.2 is trained only on real normal audio. Because MIMII-Gen is a generative model, unmodified MIMII-Gen audio may already lie outside the autoencoder's normal manifold, inflating reconstruction error and therefore AUC even without any anomaly effect. The paper does not report the AUC for unmodified MIMII-Gen audio, so the observed synthetic AUCs (e.g., 0.85 for bearings vs. 0.57 real) may reflect a domain-shift artifact rather than the LLM-selected effects. Please report the no-effect baseline for each machine type and, if needed, retrain the detector using the same MIMII-Gen normal audio as the negative class to remove this confound.
  2. [§4.3, Table 3] The rank-consistency evidence comes from a single autoencoder run and fixed sets of 50 synthetic and 50 real anomalies per machine type, with no confidence intervals, bootstrap, or permutation test. With only five machine types, the observed perfect agreement in each scoring function is suggestive but does not tell the reader how stable the ranking is to the particular anomaly samples, the autoencoder initialization, or the LLM's non-deterministic function selection; please provide uncertainty quantification (e.g., bootstrap over anomalies, repeated autoencoder training, or multiple LLM samples) to support the claim.
  3. [§5, Conclusion and §4.3] The statement that rankings are consistent across "different anomaly detection systems" is not supported by the experiments, since the MSE and Mahalanobis scores in Table 3 are two outputs of the same autoencoder representation rather than independent detection systems. Please either rephrase the claim or evaluate at least one genuinely different architecture (e.g., a GMM or an embedding-similarity method) to substantiate the cross-system claim.
  4. [§4.4, Table 4] The ablation asserts that GPT-4o and manual-mapping rankings "closely aligned" with real anomalies while random selection showed no correlation, but no quantitative correlation coefficient or significance test is reported; the random column actually produces a nearly inverted ranking (e.g., valve AUC 0.89, rank 1, vs. real rank 5). Please report Spearman/Kendall correlations with confidence intervals or permutation-based p-values, and state whether the differences among the three configurations are statistically meaningful.
minor comments (5)
  1. [§4.1] The sentence "Table 2 summarizes the datasets used in this study." appears twice in consecutive paragraphs; one occurrence should be deleted.
  2. [§3.1.2] The example add_squeaking has fixed parameters (freq=4000, intensity=0.3), but the workflow in §3.1.1 only says the LLM selects the effect function; please state explicitly whether the LLM also chooses parameter values, since the claim of "diverse and controllable" synthesis depends on this.
  3. [§4.1] The provenance of the 50 real anomalous recordings per machine type is not stated; specify whether they come from MIMII-DG or another dataset and describe the fault types and severity range.
  4. [Table 3] In the synthetic block, the "AUC Rank" column is placed before the MSE and MAHALA columns, making it unclear whether the rank refers to one metric, the other, or an average; clarify the column layout.
  5. [Eq. (1)] The function H(x) is defined to return 1 if x>0 and 0 otherwise, which treats ties as negatives; state this explicitly, since AUC implementations vary in tie handling.

Circularity Check

0 steps flagged · score 2.0 of 10

No circular reduction: the synthetic-vs-real AUC ranking agreement is an empirical result validated against real fault recordings; only minor self-dependency on the authors' MIMII-Gen/MIMII-DG resources warrants note.

full rationale

The derivation chain is: (1) Flan-T5 generates a caption from metadata; (2) GPT-4 selects DSP effect functions from a fixed library and applies them to MIMII-Gen normal audio (Sec. 3.1); (3) an autoencoder trained only on real normal audio scores both synthetic and real anomalies; (4) Table 3 compares AUC rankings. No step fits a parameter to real anomaly AUCs and then calls it a prediction: the LLM receives captions and human-written effect descriptions but no detection scores, and Eq. (1) is a standard AUC definition applied symmetrically to synthetic and real anomalies. The central claim, that synthetic and real rankings agree, is not forced by construction: Table 4 shows that with the same MIMII-Gen audio base held fixed, random effect selection destroys the agreement (Fan rank 4, Valve rank 1) while GPT-4o and manual mapping reproduce the real ranking, so the effect-to-machine mapping, not the generation backbone, drives the result. Validation is self-contained against external physical fault recordings (real anomalies from MIMII-DG). The main self-dependency is the authors' own MIMII-Gen [10] as the normal-audio backbone and MIMII-DG [21] as the data source; these are components, not justifications, of the central claim, and the cited 'high-fidelity' property of MIMII-Gen is testable here against real audio. Residual validity risks, which are not circularity: no AUC for unmodified MIMII-Gen audio is reported, so the domain-shift component of the uniformly higher synthetic AUCs (Sec. 4.3) is unquantified; the n=5 ranking has no confidence intervals; and 'different anomaly detection systems' refers to MSE and Mahalanobis scores of the same autoencoder representation, an overstatement of independence. None of these reduces the ranking result to its own inputs, so the circularity score is low.

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

The central claim rests on the realism of MIMII-Gen normal audio, the coverage of the hand-designed effect library, and the stability of AUC estimates without error bars. No genuinely new physical entities are introduced; the method depends on assumptions about synthetic data realism that are not independently verified.

free parameters (3)
  • DSP effect parameters (e.g., add_squeaking: freq=4000 Hz, intensity=0.3, duration=2.0 s) = 4000 Hz, 0.3, 2.0 s
    Chosen by hand; these settings determine the character and detectability of the synthetic anomalies and are not varied or justified.
  • Number of synthetic anomalies per machine type = 50
    Arbitrary choice; no power analysis or sensitivity study is shown.
  • Autoencoder hyperparameters (mel bins, layers, filters, learning rate, epochs) = 128 bins, 3 layers, 128/64/32 filters, Adam lr=0.001, 100 epochs
    Standard choices, but they define the single system used; results are not shown to be robust to these choices.
assumptions (3)
  • domain assumption MIMII-Gen generates normal machine sounds that are acoustically representative of real normal sounds from MIMII-DG.
    Invoked in §3.1.1 where 'Generate normal audio: Using the MIMII-Gen latent diffusion model [10]' provides the base for synthetic anomalies; if the generated normal audio differs systematically from real normal audio, the synthetic anomalies may not reflect real detection difficulty.
  • domain assumption The hand-crafted sound effect library (squeaking, rattling, grinding, etc.) covers the space of plausible machine faults.
    The effects are defined in §3.1.2 without validation that they align with real fault acoustics; the LLM can only choose among these functions.
  • domain assumption AUC computed from a single run with 50 positives and 900 negatives is stable enough for ranking.
    No confidence intervals or repeated trials are reported (§4.3, Table 3), yet the rank ordering is treated as reliable.

how reviews work

0 comments
Cite this review

Pith. "Pith review of MIMII-Agent: Leveraging LLMs with Function Calling for Relative Evaluation of Anomalous Sound Detection." pith.science (2026). https://pith.science/paper/SCMT27FB

@misc{pith2026250720666,
  author       = {Pith},
  title        = {Pith review of: MIMII-Agent: Leveraging LLMs with Function Calling for Relative Evaluation of Anomalous Sound Detection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SCMT27FB}},
  note         = {Machine review of arXiv:2507.20666}
}
read the original abstract

This paper proposes a method for generating machine-type-specific anomalies to evaluate the relative performance of unsupervised anomalous sound detection (UASD) systems across different machine types, even in the absence of real anomaly sound data. Conventional keyword-based data augmentation methods often produce unrealistic sounds due to their reliance on manually defined labels, limiting scalability as machine types and anomaly patterns diversify. Advanced audio generative models, such as MIMII-Gen, show promise but typically depend on anomalous training data, making them less effective when diverse anomalous examples are unavailable. To address these limitations, we propose a novel synthesis approach leveraging large language models (LLMs) to interpret textual descriptions of faults and automatically select audio transformation functions, converting normal machine sounds into diverse and plausible anomalous sounds. We validate this approach by evaluating a UASD system trained only on normal sounds from five machine types, using both real and synthetic anomaly data. Experimental results reveal consistent trends in relative detection difficulty across machine types between synthetic and real anomalies. This finding supports our hypothesis and highlights the effectiveness of the proposed LLM-based synthesis approach for relative evaluation of UASD systems.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

34 extracted references · 33 canonical work pages

  1. [1]

    Description and discussion on DCASE2020 challenge task2: Unsupervised anomalous sound detection for machine condition monitoring,

    Y . Koizumi, Y . Kawaguchi, K. Imoto, T. Nakamura, Y . Nikaido, R. Tanabe, H. Purohit, K. Suefusa, T. Endo, M. Yasuda, and N. Harada, “Description and discussion on DCASE2020 challenge task2: Unsupervised anomalous sound detection for machine condition monitoring,” in Proc. Workshop on Detection and Classification of Acoustic Scenes and Events (DCASE) , 2...

    work page 2020

  2. [2]

    Audio surveillance of roads: A system for detecting anomalous sounds,

    P. Foggia, N. Petkov, A. Saggese, N. Strisciuglio, and M. Vento, “Audio surveillance of roads: A system for detecting anomalous sounds,” IEEE transactions on intelligent transportation systems , vol. 17, no. 1, pp. 279–288, 2015

    work page 2015

  3. [3]

    Deep autoencoders for acoustic anomaly detection: Experi- ments with working machine and in-vehicle audio,

    G. Coelho, L. M. Matos, P. J. Pereira, A. Ferreira, A. Pilastri, and P. Cortez, “Deep autoencoders for acoustic anomaly detection: Experi- ments with working machine and in-vehicle audio,” Neural Computing and Applications, vol. 34, no. 22, pp. 19 485–19 499, 2022

    work page 2022

  4. [4]

    Anomalous sound detection based on interpolation deep neural network,

    K. Suefusa, T. Nishida, H. Purohit, R. Tanabe, T. Endo, and Y . Kawaguchi, “Anomalous sound detection based on interpolation deep neural network,” in Proc. IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2020, pp. 271–275

    work page 2020

  5. [5]

    Detection of anomalous sounds for machine condition monitoring using classification confidence,

    T. Inoue, P. Vinayavekhin, S. Morikuni, S. Wang, T. H. Trong, D. Wood, M. Tatsubori, and R. Tachibana, “Detection of anomalous sounds for machine condition monitoring using classification confidence,” in Proc. Workshop on Detection and Classification of Acoustic Scenes and Events (DCASE), 2020, pp. 66—-70

    work page 2020

  6. [6]

    Fraunhofer FKIE submission for task2: First-shot unsu- pervised anomalous sound detection for machine condition monitoring,

    K. Wilkinghoff, “Fraunhofer FKIE submission for task2: First-shot unsu- pervised anomalous sound detection for machine condition monitoring,” DCASE Challenge, Tech. Rep., 2023

    work page 2023

  7. [7]

    An effective anomalous sound detection method based on representa- tion learning with simulated anomalies,

    H. Chen, Y . Song, Z. Zhuo, Y . Zhou, Y .-H. Li, H. Xue, and I. McLoughlin, “An effective anomalous sound detection method based on representa- tion learning with simulated anomalies,” in Proc. IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , 2023

    work page 2023

  8. [8]

    Regularized contrastive masked autoencoder model for machinery anomaly detection using diffusion-based data augmentation,

    E. Zahedi, M. Saraee, F. S. Masoumi, and M. Yazdinejad, “Regularized contrastive masked autoencoder model for machinery anomaly detection using diffusion-based data augmentation,” Algorithms, vol. 16, no. 9, p. 431, 2023

    work page 2023

Show all 34 references
  1. [9]

    First-shot unsupervised anomalous sound detection with unknown anomalies estimated by metadata-assisted audio generation,

    H. Zhang, Q. Zhu, J. Guan, H. Liu, F. Xiao, J. Tian, X. Mei, X. Liu, and W. Wang, “First-shot unsupervised anomalous sound detection with unknown anomalies estimated by metadata-assisted audio generation,” in Proc. IEEE International Conference on Acoustics, Speech and Signal ...

    2024

  2. [10]

    MIMII-Gen: Generative modeling approach for simulated evaluation of anomalous sound detection system,

    H. Purohit, T. Nishida, K. Dohi, T. Endo, and Y . Kawaguchi, “MIMII-Gen: Generative modeling approach for simulated evaluation of anomalous sound detection system,” arXiv preprint arXiv:2409.18542 , 2024

    2024 arXiv

  3. [11]

    Description and discussion on DCASE 2021 challenge task 2: Unsupervised anomalous sound detection for machine condition monitoring under domain shifted conditions,

    Y . Kawaguchi, K. Imoto, Y . Koizumi, N. Harada, D. Niizumi, K. Dohi, R. Tanabe, H. Purohit, and T. Endo, “Description and discussion on DCASE 2021 challenge task 2: Unsupervised anomalous sound detection for machine condition monitoring under domain shifted conditions,” in Pr...

    2021

  4. [12]

    De- scription and discussion on DCASE 2022 challenge task 2: Unsupervised anomalous sound detection for machine condition monitoring applying domain generalization techniques,

    K. Dohi, K. Imoto, N. Harada, D. Niizumi, Y . Koizumi, T. Nishida, H. Purohit, R. Tanabe, T. Endo, M. Yamamoto, and Y . Kawaguchi, “De- scription and discussion on DCASE 2022 challenge task 2: Unsupervised anomalous sound detection for machine condition monitoring applying dom...

    2022

  5. [13]

    Description and discussion on DCASE 2023 challenge task 2: First-shot unsupervised anomalous sound detection for machine condition monitoring,

    K. Dohi, K. Imoto, N. Harada, D. Niizumi, Y . Koizumi, T. Nishida, H. Purohit, R. Tanabe, T. Endo, and Y . Kawaguchi, “Description and discussion on DCASE 2023 challenge task 2: First-shot unsupervised anomalous sound detection for machine condition monitoring,” in Proc. Works...

    2023

  6. [14]

    Description and discussion on DCASE 2024 challenge task 2: First- shot unsupervised anomalous sound detection for machine condition monitoring,

    T. Nishida, N. Harada, D. Niizumi, D. Albertini, R. Sannino, S. Pradolini, F. Augusti, K. Imoto, K. Dohi, H. Purohit, T. Endo, and Y . Kawaguchi, “Description and discussion on DCASE 2024 challenge task 2: First- shot unsupervised anomalous sound detection for machine conditio...

    2024

  7. [15]

    ToyAD- MOS: A dataset of miniature-machine operating sounds for anomalous sound detection,

    Y . Koizumi, S. Saito, N. Harada, H. Uematsu, and K. Imoto, “ToyAD- MOS: A dataset of miniature-machine operating sounds for anomalous sound detection,” in Proc. IEEE Workshop on Applications of Signal Processing to Audio and Acoustics (WASPAA) , 2019, pp. 313–317

    2019

  8. [16]

    ToyADMOS2: Another dataset of miniature-machine operating sounds for anomalous sound detection under domain shift conditions,

    N. Harada, D. Niizumi, D. Takeuchi, Y . Ohishi, M. Yasuda, and S. Saito, “ToyADMOS2: Another dataset of miniature-machine operating sounds for anomalous sound detection under domain shift conditions,” in Proc. Workshop on Detection and Classification of Acoustic Scenes and Eve...

    2021

  9. [17]

    ToyADMOS2+: New ToyADMOS data and benchmark results of the first-shot anomalous sound event detection baseline,

    N. Harada, D. Niizumi, D. Takeuchi, Y . Ohishi, and M. Yasuda, “ToyADMOS2+: New ToyADMOS data and benchmark results of the first-shot anomalous sound event detection baseline,” in Proc. Workshop on Detection and Classification of Acoustic Scenes and Events (DCASE) , 2023, pp. 41–45

    2023

  10. [18]

    ToyADMOS2: Yet another dataset for the DCASE2024 challenge task 2 first-shot anomalous sound detection,

    D. Niizumi, N. Harada, Y . Ohishi, D. Takeuchi, and M. Yasuda, “ToyADMOS2: Yet another dataset for the DCASE2024 challenge task 2 first-shot anomalous sound detection,” in Proc. Workshop on Detection and Classification of Acoustic Scenes and Events (DCASE) , 2024, pp. 106–110

    2024

  11. [19]

    MIMII dataset: Sound dataset for malfunctioning industrial machine investigation and inspection,

    H. Purohit, R. Tanabe, K. Ichige, T. Endo, Y . Nikaido, K. Suefusa, and Y . Kawaguchi, “MIMII dataset: Sound dataset for malfunctioning industrial machine investigation and inspection,” in Proc. Workshop on Detection and Classification of Acoustic Scenes and Events (DCASE) , 2...

    2019

  12. [20]

    MIMII DUE: Sound dataset for malfunctioning industrial machine investigation and inspection with domain shifts due to changes in operational and environmental conditions,

    R. Tanabe, H. Purohit, K. Dohi, T. Endo, Y . Nikaido, T. Nakamura, and Y . Kawaguchi, “MIMII DUE: Sound dataset for malfunctioning industrial machine investigation and inspection with domain shifts due to changes in operational and environmental conditions,” in Proc. IEEE Work...

    2021

  13. [21]

    MIMII DG: Sound dataset for mal- functioning industrial machine investigation and inspection for domain generalization task,

    K. Dohi, T. Nishida, H. Purohit, R. Tanabe, T. Endo, M. Yamamoto, Y . Nikaido, and Y . Kawaguchi, “MIMII DG: Sound dataset for mal- functioning industrial machine investigation and inspection for domain generalization task,” in Proc. Workshop on Detection and Classification of...

    2022

  14. [22]

    Unsupervised detection of anomalous sound based on deep learning and the Neyman- Pearson lemma,

    Y . Koizumi, S. Saito, H. Uematsu, and N. Harada, “Unsupervised detection of anomalous sound based on deep learning and the Neyman- Pearson lemma,” IEEE/ACM Transactions on Audio, Speech, and Language Processing, vol. 27, no. 1, pp. 212–224, 2019

    2019

  15. [23]

    Anomalous sound event detection: A survey of machine learning based methods and applications,

    Z. Mnasri, S. Rovetta, and F. Masulli, “Anomalous sound event detection: A survey of machine learning based methods and applications,” Multimedia Tools and Applications , vol. 81, pp. 1–35, 2021

    2021

  16. [24]

    Unsu- pervised anomalous sound detection for machine condition monitoring using classification-based methods,

    Y . Wang, Y . Zheng, Y . Zhang, Y . Xie, S. Xu, Y . Hu, and L. He, “Unsu- pervised anomalous sound detection for machine condition monitoring using classification-based methods,” Applied Sciences, vol. 11, no. 23, p. 11128, 2021

    2021

  17. [25]

    AudioLDM: Text-to-audio generation with latent diffusion models,

    H. Liu, Z. Chen, Y . Yuan, X. Mei, X. Liu, D. Mandic, W. Wang, and M. D. Plumbley, “AudioLDM: Text-to-audio generation with latent diffusion models,” in Proc. International Conference on Machine Learning (ICML), 2023, pp. 21 450–21 474

    2023

  18. [26]

    CLAP: Learning audio concepts from natural language supervision,

    B. Elizalde, S. Deshmukh, M. Al Ismail, and H. Wang, “CLAP: Learning audio concepts from natural language supervision,” in Proc. IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2023

    2023

  19. [27]

    Make-An-Audio: Text-to-audio generation with prompt-enhanced diffusion models,

    R. Huang, J. Huang, D. Yang, Y . Ren, L. Liu, M. Li, Z. Ye, J. Liu, X. Yin, and Z. Zhao, “Make-An-Audio: Text-to-audio generation with prompt-enhanced diffusion models,” in Proc. International Conference on Machine Learning (ICML) , 2023, pp. 13 916–13 932

    2023

  20. [28]

    Make-An-Audio 2: Temporal-enhanced text-to- audio generation,

    J. Huang, Y . Ren, R. Huang, D. Yang, Z. Ye, C. Zhang, J. Liu, X. Yin, Z. Ma, and Z. Zhao, “Make-An-Audio 2: Temporal-enhanced text-to- audio generation,” arXiv preprint arXiv:2305.18474 , 2023

    2023 arXiv

  21. [29]

    AudioLDM 2: Learning holistic audio generation with self-supervised pretraining,

    H. Liu, Y . Yuan, X. Liu, X. Mei, Q. Kong, Q. Tian, Y . Wang, W. Wang, Y . Wang, and M. D. Plumbley, “AudioLDM 2: Learning holistic audio generation with self-supervised pretraining,” IEEE/ACM Transactions on Audio, Speech, and Language Processing , vol. 32, pp. 2871–2883, 2024

    2024

  22. [30]

    AudioLCM: Text-to-audio generation with latent consistency models,

    H. Liu, R. Huang, Y . Liu, H. Cao, J. Wang, X. Cheng, S. Zheng, and Z. Zhao, “AudioLCM: Text-to-audio generation with latent consistency models,” in Proc. ACM International Conference on Multimedia (MM) , 2024, pp. 7008––7017

    2024

  23. [31]

    ConsistencyTTA: Accelerating diffusion-based text-to-audio generation with consistency distillation,

    Y . Bai, T. Dang, D. Tran, K. Koishida, and S. Sojoudi, “ConsistencyTTA: Accelerating diffusion-based text-to-audio generation with consistency distillation,” in Proc. INTERSPEECH, 2024, pp. 3285–3289

    2024

  24. [32]

    SoundCTM: Unifying score-based and consistency models for full-band text-to-sound generation,

    K. Saito, D. Kim, T. Shibuya, C.-H. Lai, Z. Zhong, Y . Takida, and Y . Mitsufuji, “SoundCTM: Unifying score-based and consistency models for full-band text-to-sound generation,” in Proc. International Conference on Learning Representations (ICLR) , 2025

    2025

  25. [33]

    Tango 2: Aligning diffusion-based text-to-audio generations through direct preference optimization,

    N. Majumder, C.-Y . Hung, D. Ghosal, W.-N. Hsu, R. Mihalcea, and S. Poria, “Tango 2: Aligning diffusion-based text-to-audio generations through direct preference optimization,” in Proc. ACM International Conference on Multimedia (MM) , 2024, pp. 564—-572

    2024

  26. [34]

    Scaling instruction-finetuned language models,

    H. W. Chung, L. Hou, S. Longpre, B. Zoph, Y . Tay, W. Fedus, Y . Li, X. Wang, M. Dehghani, S. Brahma et al., “Scaling instruction-finetuned language models,” Journal of Machine Learning Research , vol. 25, no. 1, pp. 70:1–70:53, 2024

    2024

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.