Pith. sign in

REVIEW 5 major objections 5 minor 2 cited by

Temporal Attention Pooling for Frequency Dynamic Convolution in Sound Event Detection

T0 review · 5 major / 5 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read Replacing temporal average pooling with time-plus-velocity attention improves sound event detection by 3% and sets a new state of the art.

desk verdict Useful incremental extension of FDY conv with a likely real gain on DESED, but the velocity-attention branch is mis-specified in the equations and the SOTA claim rests on best-of-run/best-of-configuration numbers. read the letter →

arxiv 2504.12670 v1 pith:G2PLBYD5 submitted 2025-04-17 eess.AS cs.SD

Hyeonuk Nam , Yong-Hwa Park This is my paper · ORCID
classification eess.AScs.SD
keywords soundeventdetectiontemporalattentionpoolingfrequencydynamicconvolutionvelocityPSDS1DESEDdatasettransienteventssemi-supervisedlearning
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 claims that the temporal average pooling used inside frequency dynamic convolution (FDY conv) for sound event detection is a bottleneck: it weights every frame equally and dilutes short, transient sounds such as alarms, door knocks, and speech plosives. To fix this, the authors propose temporal attention pooling (TAP), which sums three pooling branches: a saliency-weighted time attention branch, a velocity attention branch that weights frames by their temporal differences, and the original average pooling for stationary sounds. Replacing only the pooling stage raises the average polyphonic sound detection score (PSDS1) from 0.431 to 0.444 over FDY conv with a 14.8% parameter increase, and the best TAP-plus-multi-dilated configuration reaches a maximum PSDS1 of 0.459, which the paper reports as surpassing previous state-of-the-art systems on the DESED benchmark. A reader should care because the change is a drop-in swap of the pooling operation, not a new architecture, and it specifically targets the transient events that average pooling systematically under-detects.

What carries the argument

The machinery is temporal attention pooling (TAP), the three-branch pooling operator defined by $x_{\mathrm{TAP}} = \sum_{t=1}^{T}\alpha_t \odot x_{s,t} + \sum_{t=1}^{T}\beta_t \odot x_{s,t} + \frac{1}{T}\sum_{t=1}^{T}x_t$. The first term is time attention pooling: softmax attention weights $\alpha_t$ computed from 2D convolutions over the input, applied to a saliency-enhanced map $x_s$ obtained by two 2D convolutions with a sigmoid gate. The second term is velocity attention pooling: the same structure but with weights $\beta_t$ computed from the first-order temporal difference $\Delta x_t = x_t - x_{t-1}$, so frames with rapid spectral change receive higher weight. The third term is the original temporal average pooling on the raw feature, kept for robustness to stationary signals. The three outputs are summed without normalization, and the result replaces the average-pooled feature that FDY conv uses to generate frequency-adaptive kernels. The additional parameters come from the two saliency convolutions and the attention-weight convolutions, about 14.8% over FDY conv.

What would settle it

Train TFD conv with the velocity branch replaced by a boundary-handled first difference (e.g., $\Delta x_1 = x_2 - x_1$) and the three pooling outputs renormalized; if average PSDS1 on transient-heavy classes stays at or above 0.444, the specific unnormalized three-branch design is not what carries the gain. If the score instead drops below the FDY conv baseline of 0.431, the velocity branch and its scaling are load-bearing.

Watch

Extended reading notes

Core claim

The central claim is that the temporal aggregation step inside FDY conv—not the frequency-adaptive kernels themselves—limits detection of transient events, and that replacing temporal average pooling with TAP resolves this. TAP computes the pooled feature as the sum of three terms: time attention pooling, where softmax weights $\alpha_t$ are applied to a saliency-enhanced representation $x_s$ obtained from two 2D convolutions; velocity attention pooling, where softmax weights $\beta_t$ are derived from first-order temporal differences $\Delta x_t = x_t - x_{t-1}$; and plain average pooling of the raw input. In ablation, all three branches together give the best result (0.444 average PSDS1), with velocity attention alone slightly outperforming time attention alone (0.440 vs 0.439), and the three-way sum behaving better than any pair. The paper further claims that the same pooling replacement transfers across FDY variants—dilated, partial, and multi-dilated—with the best configuration, TAP plus MDFD conv, reaching 0.459 maximum PSDS1 and outperforming the baseline CRNN by a wide margin.

Load-bearing premise

The claim rests on the assumption that first-order temporal differences, with no boundary handling specified for the first frame, capture the transient salience that average pooling misses, and that simply summing the three pooling outputs yields a properly scaled feature for kernel generation.

Editorial extensions

If this is right

  • A drop-in swap of the pooling stage in FDY conv yields a 3.02% average PSDS1 improvement (0.444 vs 0.431) with only 14.8% more parameters, so temporal pooling is a cheap place to gain sound event detection performance.
  • TAP generalizes across FDY variants: it improves or matches PFD conv at reduced dynamic-channel ratios, and its best combination with MDFD conv reaches 0.459 maximum PSDS1, the highest reported in the paper.
  • Classwise statistical analysis (ANOVA with Tukey HSD) shows TFD conv significantly beats other models on transient-heavy classes such as alarm/bell, cat, dog, and electric shaver, while keeping up on quasi-stationary classes—except vacuum cleaner, where the plain CRNN remains best.
  • Dilation in the frequency dimension does not help PSDS1 when combined with TAP; all dilated TAP+DFD variants score below the non-dilated TFD conv, suggesting that aggressive spectral receptive-field expansion can smooth away the transient cues TAP preserves.

Reading between the lines

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

  • The unnormalized sum of three pooling branches leaves the overall scale of $x_{\mathrm{TAP}}$ dependent on attention-weight magnitudes; one testable extension is whether normalizing the sum (e.g., dividing by the sum of $\alpha$ and $\beta$ plus 1) changes stability across configurations or datasets.
  • Velocity attention is a first-order finite difference; second-order or multi-scale difference operators could capture acceleration of spectral change (e.g., frequency sweeps) and might extend the transient benefit beyond the classes tested here.
  • Because TAP is a generic pooling layer, the same recipe could be dropped into other average-pooling-dependent modules in audio tasks where transient cues matter, such as bioacoustic detection or audio tagging.
  • The paper reports only PSDS1; since PSDS2 emphasizes presence/absence rather than temporal boundaries, TAP's effect on audio-tagging-style evaluation remains untested and could differ.
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

5 major / 5 minor

Summary. The paper proposes temporal attention pooling frequency dynamic convolution (TFD conv), which replaces the temporal average pooling inside frequency dynamic convolution (FDY conv) with a temporal attention pooling (TAP) module. TAP consists of three branches: time attention pooling (TA), velocity attention pooling (VA), and conventional average pooling. Experiments on the DESED dataset with a CRNN backbone and mean-teacher semi-supervised training report an average PSDS1 improvement from 0.431 (FDY) to 0.444 (TFD), and a maximum PSDS1 of 0.459 for a combined TAP+MDFD configuration, which the authors describe as surpassing prior state-of-the-art SED systems. Classwise ANOVA and Tukey HSD analyses are used to support the claim that the gains concentrate on transient-heavy event classes.

Significance. If the claims hold, TAP would be a simple, drop-in modification to FDY-style SED models with modest parameter overhead and released code, which is valuable for reproducibility and for the DCASE-oriented community. However, the manuscript currently contains a load-bearing inconsistency in the definition of the velocity attention branch, an internal mismatch between the equations and the model that was actually ablated, no variance information for the reported average scores, and a state-of-the-art claim based on maximum scores over runs and configurations without external comparison. These issues must be resolved before the contribution and its mechanistic interpretation can be accepted with confidence.

major comments (5)
  1. [Section III-B2, Eqs. (6)-(8)] Equation (6) defines Δx_t = x_t − x_{t−1}, but Equation (7) computes β from x, not from Δx, and Δx never appears again. As written, the velocity attention branch does not use temporal differences at all; Eq. (6) is inert. Consequently, the Table II result 'FDY w/ VA' (0.440) demonstrates only that an additional attention branch with independently learned parameters helps, not that difference-based transient attention helps. This undermines the central mechanistic claim that TAP improves transient-event detection through velocity cues. Please either amend Eq. (7) to consume Δx, with an explicit boundary convention for t=1, or revise the description, figure, and interpretation of the VA branch to match what is actually computed.
  2. [Section III-B, Eqs. (4) and (7); Section V-A] Equations (4) and (7) compute α and β from the raw input x, but Section V-A states that 'both time attention and velocity attention branches use xs as input to compute attention weights,' and Table I shows that the reported model uses xs (0.439 vs. 0.434 for TA with x). The equations therefore do not describe the model whose results are reported. Please align the equations with the actual implementation, including the path through xs, so that the method is reproducible from the text.
  3. [Section IV-G and Tables III, VIII] All average PSDS1 values are means over 12 runs with no standard deviations or confidence intervals, so the 3.02% improvement (0.444 vs. 0.431) cannot be judged for statistical significance. The 'state-of-the-art' claim (Abstract and Section V-H) is based on the maximum over runs and over the TAP+MDFD configurations in Table VII, and Table VIII compares only the authors' own FDY variants, not published SOTA systems on DESED (e.g., transformer-based systems cited in [44]-[49]). Please report variance, use a consistent statistic for performance claims, and add an external comparison on the same evaluation protocol.
  4. [Section III-B, Eqs. (3)-(5)] The softmax axis in Eqs. (4) and (7) is not specified. Because x_TAP is formed by summing α_t⊙x_s,t and β_t⊙x_s,t over t and adding a time average, the normalization of α and β determines the scale of the pooled feature and the relative weight of the three branches. Please state the softmax dimension and any normalization applied to the sum of the three terms; otherwise the equations are ambiguous and the architecture cannot be reproduced exactly.
  5. [Section V-D, Table IV] The ANOVA/Tukey HSD results are reported only as qualitative inequality chains (e.g., 'DFD < TFD'). No test statistics, p-values, degrees of freedom, or the number of runs entering the analysis are given. As presented, the claim that TFD conv 'significantly enhances' transient-heavy classes is not substantiated. Please include the full statistical output or remove the significance language.
minor comments (5)
  1. [Section V-B] Typo: 'thaft' in 'indicating thaft while TA is useful' should be 'that'.
  2. [Abstract and throughout] Inconsistent spacing appears in 'ANOV A', 'V A', and 'TAP + MDFD convs' (Section IV-A); please use consistent notation.
  3. [Fig. 1 caption] The caption reads '(b)velocity'; a space is missing after '(b)'.
  4. [Section IV-G and Table VIII] Section IV-G says all PSDS1 scores in the tables are averages over twelve runs, but Table VIII reports 'Max PSDS1'. Please clarify which tables report averages and which report maxima.
  5. [References] Reference [56] appears to duplicate reference [47] (same title and venue); please merge or differentiate them.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the TFD conv claim is an empirical comparison on the external DESED/PSDS1 benchmark, with no fitted quantity renamed as a prediction and no load-bearing self-citation chain.

full rationale

The paper's central claim is that replacing temporal average pooling in FDY conv with TAP improves PSDS1 on the external DESED benchmark. The improvement is measured empirically across twelve training runs with ablations (Tables I, II, III), so the result is not derived from its own definition. FDY conv, DFD conv, PFD conv, and MDFD conv are prior models by the same authors, but they serve as baselines or integration targets rather than as assumptions that force the TAP result; citing one's own published baselines is normal comparison practice, not circularity. No parameter is fitted to a subset of data and then called a prediction: the reported scores come from training on benchmark splits and evaluating on the real validation set. No uniqueness theorem or external authority is invoked to forbid alternatives. The skeptical observation that Eq. (6) defines Δx_t but Eq. (7) computes β from x rather than from Δx is an internal implementation inconsistency or a possible error, not a circular reduction: the velocity attention branch as written may not implement temporal differences, and that could weaken the mechanistic interpretation of the VA branch, but it does not make the empirical claim equivalent to its inputs. The paper is self-contained against an external benchmark, so the honest circularity finding is a score of 0.

Assumptions & free parameters 1 free parameters · 4 assumptions · 0 invented entities

The paper introduces no new physical entities. Its central claim rests on learned attention modules and two architectural assumptions: that temporal differences encode transients and that summing three pooling branches is the right way to combine them. All evaluations use the existing DESED/PSDS1 setup.

free parameters (1)
  • TAP+MDFD best architecture (channel ratio 5/4, dilation (1)x3+(2,2,3)) = Average PSDS1 0.445, maximum PSDS1 0.459
    Selected as the best of 14 configurations in Table VII after search; the headline state-of-the-art number depends on this post hoc selection.
assumptions (4)
  • domain assumption The FDY conv family, including DFD, PFD, and MDFD conv, is correctly reproduced as described in references [37], [42], and [43].
    The proposed method is built on these prior models and uses them as baselines; no independent re-implementation check is included.
  • ad hoc to paper Temporal average pooling inside FDY conv dilutes transient cues, so attention-based weighting should recover them.
    The central motivation is asserted in Section III.A; the ablations support it empirically but there is no formal argument.
  • ad hoc to paper First-order temporal differences x_t - x_{t-1} are a useful proxy for transient salience.
    Used to construct velocity attention in Eq. (6); no theoretical or perceptual evidence is given.
  • domain assumption PSDS1 on DESED is an appropriate external benchmark for SED performance.
    The metric and dataset come from prior publications and are standard in the field.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Temporal Attention Pooling for Frequency Dynamic Convolution in Sound Event Detection." pith.science (2026). https://pith.science/paper/G2PLBYD5

@misc{pith2026250412670,
  author       = {Pith},
  title        = {Pith review of: Temporal Attention Pooling for Frequency Dynamic Convolution in Sound Event Detection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/G2PLBYD5}},
  note         = {Machine review of arXiv:2504.12670}
}
read the original abstract

Recent advances in deep learning, particularly frequency dynamic convolution (FDY conv), have significantly improved sound event detection (SED) by enabling frequency-adaptive feature extraction. However, FDY conv relies on temporal average pooling, which treats all temporal frames equally, limiting its ability to capture transient sound events such as alarm bells, door knocks, and speech plosives. To address this limitation, we propose temporal attention pooling frequency dynamic convolution (TFD conv) to replace temporal average pooling with temporal attention pooling (TAP). TAP adaptively weights temporal features through three complementary mechanisms: time attention pooling (TA) for emphasizing salient features, velocity attention pooling (VA) for capturing transient changes, and conventional average pooling for robustness to stationary signals. Ablation studies show that TFD conv improves average PSDS1 by 3.02% over FDY conv with only a 14.8% increase in parameter count. Classwise ANOVA and Tukey HSD analysis further demonstrate that TFD conv significantly enhances detection performance for transient-heavy events, outperforming existing FDY conv models. Notably, TFD conv achieves a maximum PSDS1 score of 0.456, surpassing previous state-of-the-art SED systems. We also explore the compatibility of TAP with other FDY conv variants, including dilated FDY conv (DFD conv), partial FDY conv (PFD conv), and multi-dilated FDY conv (MDFD conv). Among these, the integration of TAP with MDFD conv achieves the best result with a PSDS1 score of 0.459, validating the complementary strengths of temporal attention and multi-scale frequency adaptation. These findings establish TFD conv as a powerful and generalizable framework for enhancing both transient sensitivity and overall feature robustness in SED.

Figures

Figures reproduced from arXiv: 2504.12670 by the authors.

Figure 1
Figure 1. Overview of the proposed temporal attention pooling frequency dynamic convolution (TFD conv). The left side illustrates the overall archi￾tecture of the TFD conv-based SED model, where TFD conv layers replace standard FDY conv layers for enhanced time-frequency adaptive feature extraction. The right side provides a detailed breakdown of the temporal attention pooling (TAP) mechanism, which replaces temporal average … view at source ↗
Figure 2
Figure 2. Illustration of different frequency dynamic convolution (FDY conv) variants. (a) [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Overview of the proposed temporal attention pooling (TAP) mechanism. TAP consists of three pooling branches: (a) Attention pooling applies softmax-based attention to highlight salient temporal features. (b) Velocity attention pooling incorporates temporal differences (∆x) and applies softmax weighting to emphasize transient sound patterns. (c) Average pooling captures global temporal context by computing the mean ov… view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Classwise F1 score distribution of different models across ten sound [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Auditory Intelligence: Understanding the World Through Sound

    eess.AS 2025-08 conditional novelty 4.0 of 10

    A position paper proposing four cognitively inspired audio tasks (ASPIRE, SODA, AUX, AUGMENT) to push machine hearing beyond recognition toward explanation, reasoning, and interaction.

  2. Frequency Dynamic Convolutions for Sound Event Detection

    eess.AS 2025-06 conditional novelty 4.0 of 10

    A family of frequency-adaptive convolutions improves CRNN sound event detection on DESED by up to 10.98% in PSDS1, with a lighter TFD variant matching the best score.

Reference graph

Works this paper leans on

59 extracted references · 21 canonical work pages · cited by 2 Pith papers

  1. [44]

    Semi-supervised learning-based sound event detection using frequency dynamic convolution with large kernel attention for DCASE challenge 2023 task 4,

    J. W. Kim, S. W. Son, Y . Song, H. K. Kim, I. H. Song, and J. E. Lim, “Semi-supervised learning-based sound event detection using frequency dynamic convolution with large kernel attention for DCASE challenge 2023 task 4,” DCASE Challenge, Tech. Rep., 2023

    work page 2023

  2. [49]

    Prototype based masked audio model for self-supervised learning of sound event detection,

    P. Cai, Y . Song, N. Jiang, Q. Gu, and I. McLoughlin, “Prototype based masked audio model for self-supervised learning of sound event detection,” arXiv preprint arXiv:2409.17656 , 2024

    arXiv 2024

  3. [1]

    Virtanen, M

    T. Virtanen, M. D. Plumbley, and D. Ellis, Computational Analysis of Sound Scenes and Events , 1st ed. Springer Publishing Company, Incorporated, 2017, pp. 3–11, 71–77

    2017

  4. [2]

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

    N. Turpault, R. Serizel, A. P. Shah, and J. Salamon, “Sound event detection in domestic environments with weakly labeled data and soundscape synthesis,” in DCASE Workshop, 2019

    2019

  5. [3]

    Convolutional recurrent neural networks for polyphonic sound event detection,

    E. C ¸ akır, G. Parascandolo, T. Heittola, H. Huttunen, and T. Virtanen, “Convolutional recurrent neural networks for polyphonic sound event detection,” IEEE/ACM Transactions on Audio, Speech, and Language Processing, vol. 25, no. 6, pp. 1291–1303, 2017

    2017

  6. [4]

    Metrics for polyphonic sound event detection,

    A. Mesaros, T. Heittola, and T. Virtanen, “Metrics for polyphonic sound event detection,” Applied Sciences, vol. 6, no. 6, 2016

    work page 2016

  7. [5]

    A framework for the robust evaluation of sound event detection,

    C ¸ . Bilen, G. Ferroni, F. Tuveri, J. Azcarreta, and S. Krstulovi ´c, “A framework for the robust evaluation of sound event detection,” in ICASSP, 2020, pp. 61–65

    2020

  8. [6]

    Study on frequency dependent convolution methods for sound event detection,

    H. Nam, S.-H. Kim, B.-Y . Ko, D. Min, and Y .-H. Park, “Study on frequency dependent convolution methods for sound event detection,” in Proc. INTER-NOISE, 2024

    work page 2024

Show all 59 references
  1. [7]

    Jitter: Jigsaw temporal transformer for event reconstruction for self-supervised sound event detection,

    H. Nam and Y .-H. Park, “Jitter: Jigsaw temporal transformer for event reconstruction for self-supervised sound event detection,” arXiv preprint arXiv:2502.20857, 2025

    2025 arXiv

  2. [8]

    SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition,

    D. S. Park, W. Chan, Y . Zhang, C.-C. Chiu, B. Zoph, E. D. Cubuk, and Q. V . Le, “SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition,” in Proc. Interspeech, 2019

    2019

  3. [9]

    Conformer: Convolution- augmented Transformer for Speech Recognition,

    A. Gulati, J. Qin, C.-C. Chiu, N. Parmar, Y . Zhang, J. Yu, W. Han, S. Wang, Z. Zhang, Y . Wu, and R. Pang, “Conformer: Convolution- augmented Transformer for Speech Recognition,” in Proc. Interspeech, 2020

    2020

  4. [10]

    Coherence-based phonemic analysis on the ef- fect of reverberation to practical automatic speech recognition,

    H. Nam and Y .-H. Park, “Coherence-based phonemic analysis on the ef- fect of reverberation to practical automatic speech recognition,” Applied Acoustics, vol. 227, p. 110233, 2025

    2025

  5. [11]

    wav2vec 2.0: A framework for self-supervised learning of speech representations,

    A. Baevski, Y . Zhou, A. Mohamed, and M. Auli, “wav2vec 2.0: A framework for self-supervised learning of speech representations,” in Advances in Neural Information Processing Systems , 2020

    2020

  6. [12]

    Hubert: Self-supervised speech representation learning by masked prediction of hidden units,

    W.-N. Hsu, B. Bolte, Y .-H. H. Tsai, K. Lakhotia, R. Salakhutdinov, and A. Mohamed, “Hubert: Self-supervised speech representation learning by masked prediction of hidden units,” IEEE/ACM Transactions on Audio, Speech, and Language Processing , 2021

    2021

  7. [13]

    Attentive statistics pooling for deep speaker embedding,

    K. Okabe, T. Koshinaka, and K. Shinoda, “Attentive statistics pooling for deep speaker embedding,” in Proc. Interspeech, 2018

    2018

  8. [14]

    Exploring the encoding layer and loss function in end-to-end speaker and language recognition system,

    W. Cai, J. Chen, and M. Li, “Exploring the encoding layer and loss function in end-to-end speaker and language recognition system,” in Proc. Interspeech, 2018

    2018

  9. [15]

    Analysis-based optimization of temporal dynamic convolutional neural network for text-independent speaker verification,

    S.-H. Kim, H. Nam, and Y .-H. Park, “Analysis-based optimization of temporal dynamic convolutional neural network for text-independent speaker verification,” IEEE Access, vol. 11, 2023

    2023

  10. [16]

    Integrating fre- quency translational invariance in tdnns and frequency positional in- formation in 2d resnets to enhance speaker verification,

    J. Thienpondt, B. Desplanques, and K. Demuynck, “Integrating fre- quency translational invariance in tdnns and frequency positional in- formation in 2d resnets to enhance speaker verification,” in Proc. Interspeech, 2021

    2021

  11. [17]

    Convolution-based channel-frequency atten- tion for text-independent speaker verification,

    J. Li, Y . Tian, and T. Lee, “Convolution-based channel-frequency atten- tion for text-independent speaker verification,” in ICASSP, 2023

    2023

  12. [18]

    Panns: Large-scale pretrained audio neural networks for audio pattern recognition,

    Q. Kong, Y . Cao, T. Iqbal, Y . Wang, W. Wang, and M. D. Plumbley, “Panns: Large-scale pretrained audio neural networks for audio pattern recognition,” IEEE/ACM Transactions on Audio, Speech, and Language Processing, 2020

    2020

  13. [19]

    Deep learning based cough detection camera using enhanced features,

    G.-T. Lee, H. Nam, S.-H. Kim, S.-M. Choi, Y . Kim, and Y .-H. Park, “Deep learning based cough detection camera using enhanced features,” Expert Systems with Applications , vol. 206, 2022

    2022

  14. [20]

    Real-time sound recognition system for human care robot considering custom sound events,

    S.-H. Kim, H. Nam, S.-M. Choi, and Y .-H. Park, “Real-time sound recognition system for human care robot considering custom sound events,” IEEE Access, vol. 12, 2024

    2024

  15. [21]

    Ast: Audio spectrogram trans- former,

    Y . Gong, Y .-A. Chung, and J. Glass, “Ast: Audio spectrogram trans- former,” in Proc. Interspeech, 2021

    2021

  16. [22]

    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,” in ICML, 2023

    2023

  17. [23]

    Overview and evaluation of sound event localization and detection in dcase 2019,

    A. Politis, A. Mesaros, S. Adavanne, T. Heittola, and T. Virtanen, “Overview and evaluation of sound event localization and detection in dcase 2019,” IEEE/ACM Transactions on Audio, Speech, and Language Processing, vol. 29, 2020

    2019

  18. [24]

    STARSS22: A dataset of spatial recordings of real scenes with spa- tiotemporal annotations of sound events,

    A. Politis, K. Shimada, P. Sudarsanam, S. Adavanne, D. Krause, Y . Koyama, N. Takahashi, S. Takahashi, Y . Mitsufuji, and T. Virtanen, “STARSS22: A dataset of spatial recordings of real scenes with spa- tiotemporal annotations of sound events,” in DCASE Workshop, 2022

    2022

  19. [25]

    Data augmentation and squeeze-and-excitation network on multiple dimension for sound event localization and detection in real scenes,

    B.-Y . Ko, H. Nam, S.-H. Kim, D. Min, S.-D. Choi, and Y .-H. Park, “Data augmentation and squeeze-and-excitation network on multiple dimension for sound event localization and detection in real scenes,” DCASE Challenge, Tech. Rep., 2022

    2022

  20. [26]

    Automated audio captioning with recurrent neural networks,

    K. Drossos, S. Adavanne, and T. Virtanen, “Automated audio captioning with recurrent neural networks,” in IEEE Workshop on Applications of Signal Processing to Audio and Acoustics , 2017

    2017

  21. [27]

    Clotho: an audio captioning dataset,

    K. Drossos, S. Lipping, and T. Virtanen, “Clotho: an audio captioning dataset,” in ICASSP, 2020

    2020

  22. [28]

    Chatgpt caption paraphrasing and fense-based caption filtering for automated audio captioning,

    I. Choi, H. Nam, D. Min, S.-D. Choi, and Y .-H. Park, “Chatgpt caption paraphrasing and fense-based caption filtering for automated audio captioning,” DCASE Challenge, Tech. Rep., 2024

    2024

  23. [29]

    Mind the Domain Gap: a Systematic Analysis on Bioacoustic Sound Event Detection,

    J. Liang, I. Nolasco, B. Ghani, H. Phan, E. Benetos, and D. Stowell, “Mind the Domain Gap: a Systematic Analysis on Bioacoustic Sound Event Detection,” arXiv preprint arXiv:2403.18638 , 2024

    2024 arXiv

  24. [30]

    Few-shot bioacoustic event detection utilizing spectro-temporal receptive field,

    D. Min, H. Nam, and Y .-H. Park, “Few-shot bioacoustic event detection utilizing spectro-temporal receptive field,” inProc. INTER-NOISE, 2024

    2024

  25. [31]

    Prtfnet: Hrtf individual- ization for accurate spectral cues using a compact prtf,

    B.-Y . Ko, G.-T. Lee, H. Nam, and Y .-H. Park, “Prtfnet: Hrtf individual- ization for accurate spectral cues using a compact prtf,” IEEE Access , vol. 11, 2023

    2023

  26. [32]

    Filteraugment: An acoustic environmental data augmentation method,

    B.-Y . Ko, Y .-H. Park, G.-T. Lee, and H. Nam, “Filteraugment: An acoustic environmental data augmentation method,” in nternational Congress on Acoustics (ICA) , 2022

    2022

  27. [33]

    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 ICML, 2023

    2023

  28. [34]

    Audiogen: Textually guided audio generation,

    F. Kreuk, G. Synnaeve, A. Polyak, U. Singer, A. D ´efossez, J. Copet, D. Parikh, Y . Taigman, and Y . Adi, “Audiogen: Textually guided audio generation,” in International Conference on Learning Representations (ICLR), 2023

    2023

  29. [35]

    Vifs: An end-to-end variational inference for foley sound synthesis,

    J. Lee, H. Nam, and Y .-H. Park, “Vifs: An end-to-end variational inference for foley sound synthesis,” DCASE Challenge, Tech. Rep., 2023

    2023

  30. [36]

    Filteraugment: An acoustic environmental data augmentation method,

    H. Nam, S.-H. Kim, and Y .-H. Park, “Filteraugment: An acoustic environmental data augmentation method,” in ICASSP, 2022

    2022

  31. [37]

    Frequency Dynamic Convolution: Frequency-Adaptive Pattern Recognition for Sound Event Detection,

    H. Nam, S.-H. Kim, B.-Y . Ko, and Y .-H. Park, “Frequency Dynamic Convolution: Frequency-Adaptive Pattern Recognition for Sound Event Detection,” in Proc. Interspeech, 2022

    2022

  32. [38]

    Towards understanding of frequency dependence on sound event detection,

    H. Nam, S.-H. Kim, D. Min, B.-Y . Ko, and Y .-H. Park, “Towards understanding of frequency dependence on sound event detection,”arXiv preprint arXiv:2502.07208, 2025

    2025 arXiv

  33. [39]

    Application of spectro-temporal re- ceptive field on soft labeled sound event detection,

    D. Min, H. Nam, and Y .-H. Park, “Application of spectro-temporal re- ceptive field on soft labeled sound event detection,” DCASE Challenge, Tech. Rep., 2023. JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2021 11

    2023

  34. [40]

    Auditory neural response inspired sound event detection based on spectro-temporal receptive field,

    ——, “Auditory neural response inspired sound event detection based on spectro-temporal receptive field,” in DCASE Workshop, 2023

    2023

  35. [41]

    Frequency & channel attention for computationally efficient sound event detection,

    H. Nam, S.-H. Kim, D. Min, and Y .-H. Park, “Frequency & channel attention for computationally efficient sound event detection,” in DCASE Workshop, 2023

    2023

  36. [42]

    Diversifying and expanding frequency-adaptive convolution kernels for sound event detection,

    H. Nam, S.-H. Kim, D. Min, J. Lee, and Y .-H. Park, “Diversifying and expanding frequency-adaptive convolution kernels for sound event detection,” in Proc. Interspeech, 2024

    2024

  37. [43]

    Pushing the limit of sound event detec- tion with multi-dilated frequency dynamic convolution,

    H. Nam and Y .-H. Park, “Pushing the limit of sound event detec- tion with multi-dilated frequency dynamic convolution,” arXiv preprint arXiv:2406.13312, 2024

    2024 arXiv

  38. [45]

    Sound event detection with weak prediction for dcase 2023 challenge task4a,

    S. Xiao, J. Shen, A. Hu, X. Zhang, P. Zhang, and Y . Yan, “Sound event detection with weak prediction for dcase 2023 challenge task4a,” DCASE Challenge, Tech. Rep., 2023

    2023

  39. [46]

    Improving audio spectrogram transformers for sound event detection through multi- stage training,

    F. Schmid, P. Primus, T. Morocutti, J. Greif, and G. Widmer, “Improving audio spectrogram transformers for sound event detection through multi- stage training,” DCASE2024 Challenge, Tech. Rep., 2024

    2024

  40. [47]

    Self training and ensembling frequency dependent networks with coarse prediction pooling and sound event bounding boxes,

    H. Nam, D. Min, I. Choi, S.-D. Choi, and Y .-H. Park, “Self training and ensembling frequency dependent networks with coarse prediction pooling and sound event bounding boxes,” in DCASE Workshop, 2024

    2024

  41. [48]

    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,” in Proc. Interspeech, 2024

    2024

  42. [50]

    Convolution-augmented transformer for semi-supervised sound event detection,

    K. Miyazaki, T. Komatsu, T. Hayashi, S. Watanabe, T. Toda, and K. Takeda, “Convolution-augmented transformer for semi-supervised sound event detection,” DCASE Challenge, Tech. Rep., 2020

    2020

  43. [51]

    Zheng ustc team’s submission for dcase2021 task4 – semi-supervised sound event detection,

    X. Zheng, H. Chen, and Y . Song, “Zheng ustc team’s submission for dcase2021 task4 – semi-supervised sound event detection,” DCASE Challenge, Tech. Rep., 2021

    2021

  44. [52]

    Heavily augmented sound event detection utilizing weak predictions,

    H. Nam, B.-Y . Ko, G.-T. Lee, S.-H. Kim, W.-H. Jung, S.-M. Choi, and Y .-H. Park, “Heavily augmented sound event detection utilizing weak predictions,” DCASE Challenge, Tech. Rep., 2021

    2021

  45. [53]

    Frequency dependent sound event detection for dcase 2022 challenge task 4,

    H. Nam, S.-H. Kim, D. Min, B.-Y . Ko, S.-D. Choi, and Y .-H. Park, “Frequency dependent sound event detection for dcase 2022 challenge task 4,” DCASE Challenge, Tech. Rep., 2022

    2022

  46. [54]

    Multi-dimensional frequency dynamic convolution with confident mean teacher for sound event detection,

    S. Xiao, X. Zhang, and P. Zhang, “Multi-dimensional frequency dynamic convolution with confident mean teacher for sound event detection,” in ICASSP, 2023

    2023

  47. [55]

    How information on soft labels and hard labels mutually benefits sound event detection tasks,

    H. Yin, J. Bai, S. Huang, and J. Chen, “How information on soft labels and hard labels mutually benefits sound event detection tasks,” DCASE Challenge, Tech. Rep., 2023

    2023

  48. [56]

    Self training and ensembling frequency dependent networks with coarse prediction pooling and sound event bounding boxes,

    H. Nam, D. Min, I. Choi, S.-D. Choi, and Y .-H. Park, “Self training and ensembling frequency dependent networks with coarse prediction pooling and sound event bounding boxes,” DCASE Challenge, Tech. Rep., 2024

    2024

  49. [57]

    Mean teachers are better role mod- els: Weight-averaged consistency targets improve semi-supervised deep learning results,

    A. Tarvainen and H. Valpola, “Mean teachers are better role mod- els: Weight-averaged consistency targets improve semi-supervised deep learning results,” in Advances in Neural Information Processing Systems, vol. 30, 2017

    2017

  50. [58]

    mixup: Beyond empirical risk minimization,

    H. Zhang, M. Cisse, Y . N. Dauphin, and D. Lopez-Paz, “mixup: Beyond empirical risk minimization,” in International Conference on Learning Representations (ICLR), 2018

    2018

  51. [59]

    Sound event bounding boxes,

    J. Ebbers, F. G. Germain, G. Wichern, and J. L. Roux, “Sound event bounding boxes,” in Proc. Interspeech, 2024. Hyeonuk Nam received B.S. and M.S. degrees in mechanical engineering from Korea Advanced Insti- tute of Science and Technology, Daejeon, Korea, in 2018 and 2020 resp...

    2024

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.