Pith. sign in

REVIEW 3 major objections 5 minor 42 references

Sound Event Detection with Boundary-Aware Optimization and Inference

T0 review · 3 major / 5 minor · reviewed 2026-08-03 · deepseek-v4-flash

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

Pith's one-line read Modeling event boundaries directly, not just frame presence, improves sound event detection and removes the need for post-processing tuning.

desk verdict A genuinely useful boundary-aware SED method with consistent gains on the 10-class subset, but the full-AudioSet SOTA rests on hyperparameter transfer from a different setup and on single runs—worth a serious referee, conditional on sensitivity analysis and code. read the letter →

arxiv 2601.04178 v2 pith:B37PCDLI submitted 2026-01-07 eess.AS cs.SD

Florian Schmid , Chi Ian Tang , Sanjeel Parekh , Vamsi Krishna Ithapu , Juan Azcarreta Ortiz , Giacomo Ferroni , Yijun Qian , Arnoldas Jasonas
show 4 more authors
Cosmin Frateanu Camilla Clark Gerhard Widmer Çağdaş Bilen
This is my paper · ORCID
classification eess.AScs.SD
keywords soundeventdetectiontemporalboundariesproposalnetworksonset/offsetmodelingPSDS1AudioSetStrongpost-processing-freeinferencerecurrent
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

The paper tries to establish that sound event detection improves when a model is trained to predict event onsets and offsets explicitly rather than only frame-level event presence. It introduces a Recurrent Event Detection layer that converts conditional start/end probabilities into onset, offset, and presence probabilities, and Event Proposal Networks that learn to propose whole event regions with duration estimates. With tailored focal and IoU losses, the method outperforms frame-wise models with established post-processing (median filtering, SEBB, HSM3) on a 10-class subset of AudioSet Strong, and scales to all 447 classes with a PSDS1 score of 49.6, surpassing the previous best of 46.5. If correct, this means temporal localization can be improved without the separate, validation-set-dependent post-processing step that current systems rely on.

What carries the argument

The Recurrent Event Detection (RED) layer is a parameter-less differentiable recurrence that computes event presence, onset, and offset probabilities from per-frame conditional start/end probabilities, making boundaries directly trainable. The Event Proposal Networks (EPNs) are two-layer bidirectional GRUs that read RED probabilities and output two duration estimates per frame—time since onset and time to next offset—trained with an IoU loss that weights each event equally. An inference algorithm performs non-maximum suppression on these proposals, selecting up to k events per recording among the m most active classes. Together they replace non-differentiable post-processing with an end-to-e

What would settle it

Construct a held-out validation split from the AudioSet Strong training set (stratified by class), tune λiou and the k/m inference parameters on it, then evaluate on the same test set; if the PSDS1 score falls below 46.5, the reported state-of-the-art is configuration-bound rather than robust.

Watch

Extended reading notes

Core claim

The central claim is that boundary-aware optimization and inference—making onsets and offsets trainable targets and learning event proposals directly—yields more precise temporal event detection than frame-wise classification with post-processing. On a 10-class AudioSet Strong subset, the method consistently outperforms SEBB and HSM3 across CRNN, MobileNetV3+GRU, BEATs, and ATST-F backbones; the CRNN with the proposed additions matches transformer models using median filtering at roughly 60x fewer parameters. On the full AudioSet Strong, it achieves a PSDS1 score of 49.6, surpassing the previous best of 46.5 from a knowledge-distillation ensemble. The authors attribute the gains to the extra

Load-bearing premise

The full-AudioSet result assumes that configuration choices made on a 10-class subset—particularly the IoU-loss weight λiou, the inference parameters k and m, and the 70-epoch training schedule—transfer to all 447 classes, since AudioSet Strong has no predefined validation split for tuning them.

Editorial extensions

If this is right

  • Post-processing hyperparameter tuning can be removed entirely: the EPNs learn event regions during training, so no validation-set search over median filter lengths or SEBB thresholds is needed.
  • The method transfers across architectures: it improves CRNN, MobileNetV3+GRU, BEATs, and ATST-F backbones on the 10-class subset, and scales to all 447 AudioSet Strong classes.
  • A compact model can approach large-model performance: the CRNN with the proposed method (1.4M parameters) matches transformer models (~90M parameters) that use median filtering, a 60x parameter reduction from better temporal modeling.
  • The reported 49.6 PSDS1 on AudioSet Strong surpasses the previous best of 46.5, which came from a 15-model knowledge distillation ensemble, suggesting the method also avoids the complexity of ensemble distillation.
  • The boundary losses act as regularization, reducing overfitting enough that training can be shortened to 70 epochs on the full dataset.

Reading between the lines

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

  • Because the paper frames the method for temporal detection generally, the same RED+EPN machinery could plausibly transfer to other tasks with event boundaries, such as activity recognition or time-series change-point detection.
  • A natural testable extension is whether per-class GRUs, used on the 10-class subset, could be made scalable to 447 classes via grouped or hierarchical sharing, potentially improving rare-class localization beyond the single-GRU variant.
  • The authors note real-time inference is a current limitation; the RED layer's parallelizable scan and the k/m inference parameters suggest a path toward streaming SED with bounded latency.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper proposes a boundary-aware method for sound event detection (SED). It introduces the Recurrent Event Detection (RED) layer, which computes frame-wise presence, onset, and offset probabilities through a parameter-free probabilistic recurrence; an onset/offset focal loss; Event Proposal Networks (EPNs) that predict time-since-onset and time-to-offset duration estimates; and a non-maximum-suppression-style inference algorithm. The method is designed to replace post-processing such as median filtering, SEBB, and HSM3. Experiments are conducted on the AudioSet Strong subset (10 classes) and the full 447-class AudioSet Strong. On the 10-class subset, the method consistently outperforms frame-wise baselines with MF, SEBB, and HSM3 across CRNN, MN-GRU, BEATs, and ATST-F. On the full set, it reports a PSDS1 of 49.6 with BEATs, surpassing the previous best of 46.5, and also outperforms a knowledge-distillation pipeline. The paper claims that the method removes post-processing hyperparameter tuning and scales to state-of-the-art performance.

Significance. If the results hold, the paper makes a strong practical contribution: a lightweight, architecture-agnostic mechanism for temporal boundary modeling that improves SED performance across very different frame-wise backbones and eliminates tedious post-processing tuning. The RED recurrence is principled and the supervision of onset/offset probabilities and duration-based region proposals is grounded in ground-truth boundaries rather than in the evaluation metric, so there is no apparent circularity. The consistent gains on the 10-class subset across four architectures, including a 60x parameter reduction argument for CRNN, are valuable evidence. The reported full-AudioSet PSDS1 improvement is substantial. However, the headline full-set result rests on hyperparameter choices made on the 10-class subset without a validation split on the full data, and all results are single runs without variance estimates. These issues currently weaken the robustness of the central SOTA claim, although they are addressable with additional experiments and analysis.

major comments (3)
  1. [III-B, IV-C] The full-AudioSet state-of-the-art claim (Table III: 49.6 vs. 46.5) depends on hyperparameters that are not validated on the full 447-class distribution. Section III-B states that λiou is tuned on the AS-Strong-10 validation set using the per-class GRU variant, while on AS-Strong-Full the paper switches to the Single GRU variant and trains for 70 epochs 'as our method substantially reduces overfitting' — with no full validation split and no sensitivity analysis. Since the 10-class subset is short-duration, well-populated classes and the full set is long-tailed, the optimal λiou, training length, and EPN architecture may differ. Please report PSDS1 on AS-Strong-Full for a range of λiou values, training epochs, and EPN variants, or otherwise justify that the transferred configuration is robust. Without this, the 49.6 result could be a configuration artifact.
  2. [IV, Tables I–III] No error bars or repeated-seed experiments are reported for any result. The headline differences (e.g., BEATs 49.6 vs. 46.5 in Table III; CRNN 48.0 vs. 41.1 in Table I) are plausibly significant, but PSDS1 and F1 can vary with initialization and training noise, especially for transformer backbones. Please provide mean and standard deviation over at least three seeds for the main comparisons, or equivalent confidence intervals, for both AS-Strong-10 and AS-Strong-Full.
  3. [II-D, IV] The inference algorithm has two parameters, k=15 and m=|C|, which are fixed by default and not analyzed. Section II-D acknowledges that lower values reduce runtime at the potential cost of performance, but no sensitivity study is reported. The headline full-set result uses these fixed values on a different class distribution than the 10-class subset where any tuning might have occurred. Please report how PSDS1 varies with k and m on at least the 10-class validation set, and discuss whether the chosen defaults are robust for the full set. This is directly relevant to the claim that the method removes post-processing hyperparameter tuning.
minor comments (5)
  1. [Abstract] The abstract says 'using a subset' and later 'all AudioSet Strong classes'; please clarify early that the primary evaluation is on a 10-class subset and that the full-set result is a separate scaling experiment.
  2. [II-B, Eq. (3)] The notation \hat p^{on,off}_{c,t} is overloaded for both onset and offset probabilities. Defining separate symbols, e.g., \hat p^{on}_{c,t} and \hat p^{off}_{c,t}, would improve readability.
  3. [IV-B, Table II] In the configuration study, adding RED alone to CRNN decreases cF1 from 32.3 to 30.4, while adding OOL and EPN eventually improves it to 40.2. This non-monotonic behavior is not discussed; a brief explanation would help readers understand the role of the RED layer.
  4. [II-C, Eq. (4)] The phrase 'weighting by event duration ensures equal loss contribution from all events' is slightly imprecise: the denominator d^on + d^off normalizes per-frame contributions so that a long event contributes its average per-frame loss, not its total loss. Please rephrase to avoid confusion.
  5. [III-A] The description of the AS-Strong split would benefit from stating that the evaluation set of 16,935 clips is the standard evaluation portion, and from clarifying whether any clips are used for early stopping in the full-set experiments, given that no validation split exists.

Circularity Check

0 steps flagged · score 1.0 of 10

No load-bearing circular step: RED/EPN are trained on ground-truth boundaries and scored with external PSDS1; self-citations are attribution or prior baseline only.

full rationale

The derivation chain is self-contained. Equation (1) is a probabilistic recurrence relating presence to conditional start/end probabilities, and Eq. (2) defines onset/offset probabilities from the same quantities; neither embeds the evaluation metric. The losses in Eqs. (3)-(5) are supervised with ground-truth onset/offset labels, ground-truth durations, and ground-truth intervals, so the optimization target is annotated events rather than PSDS1. Algorithm 1 is a deterministic non-maximum-suppression procedure over learned duration proposals, not a fitted surrogate for the metric. The only notable self-citations are the RED patent (footnote 1, ref. [31]), which is a priority attribution and is not used to justify any experimental result, and ref. [10], which supplies the previous SOTA number (46.5) and the training/evaluation recipe; the latter is an externally published baseline, not an assumption that entails the conclusion. Transferring lambda_iou from AS-Strong-10 to AS-Strong-Full and fixing k=15, m=|C| are hyperparameter choices that raise robustness questions, but they do not make the reported 49.6 PSDS1 equal to a fitted input: the score is measured against AudioSet Strong annotations. No predicted quantity reduces by construction to a training target or to an unverified self-citation.

Assumptions & free parameters 4 free parameters · 6 assumptions · 0 invented entities

The method adds no new physical entities; its central claim rests on a parameterless recurrence, hand-set loss weights, and the assumption that fixed inference defaults generalize to the full dataset. The main fitted quantity is λiou, tuned on a 10-class validation split.

free parameters (4)
  • λiou = tuned in {0.5, 1.0, 2.0, 4.0}
    Loss weight for the IoU duration loss; tuned per model on the AS-Strong-10 validation split (Section III-B).
  • λool = 100
    Hand-fixed weight for the onset/offset focal loss; not tuned (Section II-B).
  • k and m = k=15, m=|C|
    Inference algorithm parameters fixed by default; lower values trade runtime for performance (Section II-D).
  • focal loss α = 2
    Exponent in the focal loss, fixed to 2 following prior work (Section II-B).
assumptions (6)
  • standard math Eq. (1) uses the law of total probability to decompose presence into onset plus continuation.
    The recurrence is a probabilistic identity, not an empirical assumption.
  • domain assumption The acoustic model's two sigmoid outputs are treated as conditional start/end probabilities P(e_t|¬e_{t-1}) and P(¬e_t|e_{t-1}).
    No calibration mechanism is enforced; the recurrence propagates errors from uncalibrated conditionals.
  • domain assumption AS-Strong strong annotations provide reliable frame-wise onset/offset labels.
    Training relies on exact boundary times from human annotations (Section III-A).
  • domain assumption Softplus duration estimates and the IoU loss are a suitable differentiable surrogate for boundary accuracy.
    The EPN is optimized with Eq. (4); this presumes IoU on predicted intervals correlates with PSDS1/F1.
  • ad hoc to paper Fixed inference parameters and subset-tuned λiou transfer to the full 447-class AudioSet Strong without a validation split.
    Sections II-D and III-B: defaults k=15, m=|C| and λiou tuned on AS-Strong-10 are used for AS-Strong-Full; no full validation set exists.
  • domain assumption PSDS1 with variance penalty omitted is the accepted comparison metric.
    The paper follows [10] and [35] in omitting the variance penalty; this affects the comparability of the headline SOTA.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Sound Event Detection with Boundary-Aware Optimization and Inference." pith.science (2026). https://pith.science/paper/B37PCDLI

@misc{pith2026260104178,
  author       = {Pith},
  title        = {Pith review of: Sound Event Detection with Boundary-Aware Optimization and Inference},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/B37PCDLI}},
  note         = {Machine review of arXiv:2601.04178}
}
read the original abstract

Temporal detection problems appear in many fields including time-series estimation, activity recognition and sound event detection (SED). In this work, we propose a new approach to temporal event modeling by explicitly modeling event onsets and offsets, and by introducing boundary-aware optimization and inference strategies that substantially enhance temporal event detection. The presented methodology incorporates new temporal modeling layers - Recurrent Event Detection (RED) and Event Proposal Network (EPN) - which, together with tailored loss functions, enable more effective and precise temporal event detection. We evaluate the proposed method in the SED domain using a subset of the temporally-strongly annotated portion of AudioSet. Experimental results show that our approach not only outperforms traditional frame-wise SED models with state-of-the-art post-processing, but also removes the need for post-processing hyperparameter tuning, and scales to achieve new state-of-the-art performance across all AudioSet Strong classes.

Figures

Figures reproduced from arXiv: 2601.04178 by the authors.

Figure 1
Figure 1. illustrates the overall system, its outputs, and their connections to the various loss functions. The RED layer (Section II-A), placed atop any frame-wise acoustic model, converts conditional event start and end probabilities into onset, offset, and event presence probabilities, enabling direct training on onsets and offsets (Section II-B). Rather than relying on post-processing to convert presence probabilities int… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

42 extracted references · 3 linked inside Pith

  1. [1]

    Monitoring activities of daily living in smart homes: Understanding human behavior,

    C. Debes, A. Merentitis, S. Sukhanov, M. E. Niessen, N. Frangiadakis, and A. Bauer, “Monitoring activities of daily living in smart homes: Understanding human behavior,”IEEE Signal Process. Mag., vol. 33, no. 2, pp. 81–94, 2016

    2016

  2. [2]

    A method for automatic fall detection of elderly people using floor vibrations and sound - proof of concept on human mimicking doll falls,

    Y . Zigel, D. Litvak, and I. Gannot, “A method for automatic fall detection of elderly people using floor vibrations and sound - proof of concept on human mimicking doll falls,”IEEE Trans. Biomed. Eng., vol. 56, no. 12, pp. 2858–2867, 2009

    2009

  3. [3]

    Audio analysis for surveillance applications,

    R. Radhakrishnan, A. Divakaran, and A. Smaragdis, “Audio analysis for surveillance applications,” inIEEE Workshop on Applications of Signal Processing to Audio and Acoustics. IEEE, 2005, pp. 158–161

    2005

  4. [4]

    Virtanen, M

    T. Virtanen, M. D. Plumbley, and D. Ellis,Computational Analysis of Sound Scenes and Events. Springer, 2018, vol. 9

    2018

  5. [5]

    Sound event detection: A tutorial,

    A. Mesaros, T. Heittola, T. Virtanen, and M. D. Plumbley, “Sound event detection: A tutorial,”IEEE Signal Process. Mag., vol. 38, no. 5, pp. 67–83, 2021

    2021

  6. [6]

    Event-based modeling and processing of digital media,

    R. Singh, Z. Li, P. Kim, D. Pack, and R. C. Jain, “Event-based modeling and processing of digital media,” inProceedings of the First International Workshop on Computer Vision meets Databases, vol. 66. ACM, 2004, pp. 19–26

    2004

  7. [7]

    You only hear once: A yolo-like algorithm for audio segmentation and sound event detection,

    S. Venkatesh, D. Moffat, and E. R. Miranda, “You only hear once: A yolo-like algorithm for audio segmentation and sound event detection,” Applied Sciences, vol. 12, no. 7, p. 3293, 2022

    2022

  8. [8]

    Sound event detection transformer: An event-based end-to-end model for sound event detection,

    Z. Ye, X. Wang, H. Liu, Y . Qian, R. Tao, L. Yan, and K. Ouchi, “Sound event detection transformer: An event-based end-to-end model for sound event detection,”CoRR, vol. abs/2110.02011, 2021

    arXiv 2021

Show all 42 references
  1. [9]

    Diffsed: Sound event detection with denoising diffusion,

    S. Bhosale, S. Nag, D. Kanojia, J. Deng, and X. Zhu, “Diffsed: Sound event detection with denoising diffusion,” inProceedings of the Thirty- Eighth AAAI Conference on Artificial Intelligence. AAAI Press, 2024, pp. 792–800

    2024

  2. [10]

    Effective pre-training of audio transformers for sound event detection,

    F. Schmid, T. Morocutti, F. Foscarin, J. Schl ¨uter, P. Primus, and G. Widmer, “Effective pre-training of audio transformers for sound event detection,” inIEEE International Conference on Acoustics, Speech and Signal Processing. IEEE, 2025, pp. 1–5

    2025

  3. [11]

    MAT-SED: A masked audio transformer with masked-reconstruction based pre-training for sound event detection,

    P. Cai, Y . Song, K. Li, H. Song, and I. McLoughlin, “MAT-SED: A masked audio transformer with masked-reconstruction based pre-training for sound event detection,” inAnnual Conference of the International Speech Communication Association. ISCA, 2024, pp. 557–561

    2024

  4. [12]

    Frequency dynamic convolution: Frequency-adaptive pattern recognition for sound event detection,

    H. Nam, S. Kim, B. Ko, and Y . Park, “Frequency dynamic convolution: Frequency-adaptive pattern recognition for sound event detection,” inAn- nual Conference of the International Speech Communication Association. ISCA, 2022, pp. 2763–2767

    2022

  5. [13]

    Multi- iteration multi-stage fine-tuning of transformers for sound event detection with heterogeneous datasets,

    F. Schmid, P. Primus, T. Morocutti, J. Greif, and G. Widmer, “Multi- iteration multi-stage fine-tuning of transformers for sound event detection with heterogeneous datasets,” inProceedings of the Detection and Classification of Acoustic Scenes and Events Workshop, 2024, pp. 141– 145

    2024

  6. [14]

    Audio event detection using weakly labeled data,

    A. Kumar and B. Raj, “Audio event detection using weakly labeled data,” inProceedings of the ACM Conference on Multimedia Conference. ACM, 2016, pp. 1038–1047

    2016

  7. [15]

    A closer look at weak label learning for audio events,

    A. Shah, A. Kumar, A. G. Hauptmann, and B. Raj, “A closer look at weak label learning for audio events,”CoRR, vol. abs/1804.09288, 2018

    2018 arXiv

  8. [16]

    A comparison of five multiple instance learning pooling functions for sound event detection with weak labeling,

    Y . Wang, J. Li, and F. Metze, “A comparison of five multiple instance learning pooling functions for sound event detection with weak labeling,” inIEEE International Conference on Acoustics, Speech and Signal Processing. IEEE, 2019, pp. 31–35

    2019

  9. [17]

    Convolutional recurrent neural networks for polyphonic sound event detection,

    E. C ¸akir, G. Parascandolo, T. Heittola, H. Huttunen, and T. Virtanen, “Convolutional recurrent neural networks for polyphonic sound event detection,”IEEE ACM Trans. Audio Speech Lang. Process., vol. 25, no. 6, pp. 1291–1303, 2017

    2017

  10. [18]

    Large-scale weakly supervised audio classification using gated convolutional neural network,

    Y . Xu, Q. Kong, W. Wang, and M. D. Plumbley, “Large-scale weakly supervised audio classification using gated convolutional neural network,” inIEEE International Conference on Acoustics, Speech and Signal Processing. IEEE, 2018, pp. 121–125

    2018

  11. [19]

    Sound event detection via dilated convolutional recurrent neural networks,

    Y . Li, M. Liu, K. Drossos, and T. Virtanen, “Sound event detection via dilated convolutional recurrent neural networks,” inIEEE International Conference on Acoustics, Speech and Signal Processing. IEEE, 2020, pp. 286–290

    2020

  12. [20]

    Fine-tune the pretrained ATST model for sound event detection,

    N. Shao, X. Li, and X. Li, “Fine-tune the pretrained ATST model for sound event detection,” inIEEE International Conference on Acoustics, Speech and Signal Processing. IEEE, 2024, pp. 911–915

    2024

  13. [21]

    Sound event detection in domestic environments with weakly labeled data and soundscape synthesis,

    N. Turpault, R. Serizel, J. Salamon, and A. P. Shah, “Sound event detection in domestic environments with weakly labeled data and soundscape synthesis,” inProceedings of the Detection and Classification of Acoustic Scenes and Events Workshop, 2019, pp. 253–257

    2019

  14. [22]

    Sound event bounding boxes,

    J. Ebbers, F. G. Germain, G. Wichern, and J. L. Roux, “Sound event bounding boxes,” inAnnual Conference of the International Speech Communication Association. ISCA, 2024, pp. 562–566

    2024

  15. [23]

    Onset- and-offset-aware sound event detection via differentiable frame-to-event mapping,

    T. Yoshinaga, K. Tanaka, Y . Bando, K. Imoto, and S. Morishima, “Onset- and-offset-aware sound event detection via differentiable frame-to-event mapping,”IEEE Signal Process. Lett., vol. 32, pp. 186–190, 2025

    2025

  16. [24]

    Rich feature hierarchies for accurate object detection and semantic segmentation,

    R. B. Girshick, J. Donahue, T. Darrell, and J. Malik, “Rich feature hierarchies for accurate object detection and semantic segmentation,” in IEEE Conference on Computer Vision and Pattern Recognition. IEEE Computer Society, 2014, pp. 580–587

    2014

  17. [25]

    Fast R-CNN,

    R. B. Girshick, “Fast R-CNN,” inIEEE International Conference on Computer Vision. IEEE Computer Society, 2015, pp. 1440–1448

    2015

  18. [26]

    Faster R-CNN: towards real-time object detection with region proposal networks,

    S. Ren, K. He, R. B. Girshick, and J. Sun, “Faster R-CNN: towards real-time object detection with region proposal networks,” inAdvances in Neural Information Processing Systems: Annual Conference on Neural Information Processing Systems, 2015, pp. 91–99

    2015

  19. [27]

    The hungarian method for the assignment problem,

    H. W. Kuhn, “The hungarian method for the assignment problem,” in 50 Years of Integer Programming 1958-2008 - From the Early Years to the State-of-the-Art. Springer, 2010, pp. 29–47

    1958

  20. [28]

    The benefit of temporally-strong labels in audio event classification,

    S. Hershey, D. P. W. Ellis, E. Fonseca, A. Jansen, C. Liu, R. C. Moore, and M. Plakal, “The benefit of temporally-strong labels in audio event classification,” inIEEE International Conference on Acoustics, Speech and Signal Processing. IEEE, 2021, pp. 366–370

    2021

  21. [29]

    A framework for the robust evaluation of sound event detection,

    C ¸. Bilen, G. Ferroni, F. Tuveri, J. Azcarreta, and S. Krstulovic, “A framework for the robust evaluation of sound event detection,” inIEEE International Conference on Acoustics, Speech and Signal Processing. IEEE, 2020, pp. 61–65

    2020

  22. [30]

    Threshold independent evaluation of sound event detection scores,

    J. Ebbers, R. Haeb-Umbach, and R. Serizel, “Threshold independent evaluation of sound event detection scores,” inIEEE International Conference on Acoustics, Speech and Signal Processing. IEEE, 2022, pp. 1021–1025

    2022

  23. [31]

    Sound event detection,

    C ¸. Bilen, G. Ferroni, J. A. Ortiz, F. Tuveri, and S. Krstulovic, “Sound event detection,” Patent Application Publication US20 230 317 102A1, 10 5, 2023, filed: April 5, 2022. [Online]. Available: https: //patents.google.com/patent/US20230317102A1/en

    2023

  24. [32]

    Efficient parallelization of an ubiquitous sequential computation,

    F. A. Heinsen, “Efficient parallelization of an ubiquitous sequential computation,”CoRR, vol. abs/2311.06281, 2023

    2023 arXiv

  25. [33]

    Focal loss for dense object detection,

    T. Lin, P. Goyal, R. B. Girshick, K. He, and P. Doll ´ar, “Focal loss for dense object detection,” inIEEE International Conference on Computer Vision. IEEE Computer Society, 2017, pp. 2999–3007

    2017

  26. [34]

    Dcase 2024 task 4: Sound event detection with heterogeneous data and missing labels,

    S. Cornell, J. Ebbers, C. Douwes, I. Mart´ın-Morat´o, M. Harju, A. Mesaros, and R. Serizel, “Dcase 2024 task 4: Sound event detection with heterogeneous data and missing labels,” inProceedings of the Detection and Classification of Acoustic Scenes and Events Workshop, 2024, pp. 31–35

    2024

  27. [35]

    Self-supervised audio teacher-student transformer for both clip-level and frame-level tasks,

    X. Li, N. Shao, and X. Li, “Self-supervised audio teacher-student transformer for both clip-level and frame-level tasks,”IEEE ACM Trans. Audio Speech Lang. Process., vol. 32, pp. 1336–1351, 2024

    2024

  28. [36]

    Beats: Audio pre-training with acoustic tokenizers,

    S. Chen, Y . Wu, C. Wang, S. Liu, D. Tompkins, Z. Chen, W. Che, X. Yu, and F. Wei, “Beats: Audio pre-training with acoustic tokenizers,” inInternational Conference on Machine Learning, ser. Proceedings of Machine Learning Research, vol. 202. PMLR, 2023, pp. 5178–5193

    2023

  29. [37]

    Ex- ploring performance-complexity trade-offs in sound event detection,

    T. Morocutti, F. Schmid, J. Greif, F. Foscarin, and G. Widmer, “Ex- ploring performance-complexity trade-offs in sound event detection,” in Proceedings of the European Signal Processing Conference. EURASIP, 2025

    2025

  30. [38]

    Audio set: An ontology and human-labeled dataset for audio events,

    J. F. Gemmeke, D. P. W. Ellis, D. Freedman, A. Jansen, W. Lawrence, R. C. Moore, M. Plakal, and M. Ritter, “Audio set: An ontology and human-labeled dataset for audio events,” inIEEE International Conference on Acoustics, Speech and Signal Processing. IEEE, 2017, pp. 776–780

    2017

  31. [39]

    Decoupled weight decay regularization,

    I. Loshchilov and F. Hutter, “Decoupled weight decay regularization,” in International Conference on Learning Representations. OpenReview.net, 2019

    2019

  32. [40]

    Domain generalization with relaxed instance frequency-wise normalization for multi-device acoustic scene classification,

    B. Kim, S. Yang, J. Kim, H. Park, J. Lee, and S. Chang, “Domain generalization with relaxed instance frequency-wise normalization for multi-device acoustic scene classification,” inAnnual Conference of the International Speech Communication Association. ISCA, 2022, pp. 2393–2397

    2022

  33. [41]

    Knowledge distil- lation from transformers for low-complexity acoustic scene classification,

    F. Schmid, S. Masoudian, K. Koutini, and G. Widmer, “Knowledge distil- lation from transformers for low-complexity acoustic scene classification,” inProceedings of the Detection and Classification of Acoustic Scenes and Events Workshop. Tampere University, 2022

    2022

  34. [42]

    Filteraugment: An acoustic environmental data augmentation method,

    H. Nam, S. Kim, and Y . Park, “Filteraugment: An acoustic environmental data augmentation method,” inIEEE International Conference on Acoustics, Speech and Signal Processing. IEEE, 2022, pp. 4308–4312

    2022

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.