---
title: "AudioSet / ESC-50: tasks, downloads and usage conditions"
description: "AudioSet / ESC-50: compare tasks, official access, usage conditions and evaluation. Includes research-paper mention counts and trends."
url: https://scholar4ai.com/datasets/audioset-esc-50
markdown_url: https://scholar4ai.com/datasets/audioset-esc-50.md
license: CC BY-NC 4.0 (https://creativecommons.org/licenses/by-nc/4.0/)
share_text: "AudioSet / ESC-50: compare tasks, official access, usage conditions and evaluation. Includes research-paper mention counts and trends. (as of 2026-11-05)"
data_url: https://scholar4ai.com/data/rankings.json
updated: 2026-11-05T06:02:37+09:00
window: 2026-10-07 to 2026-11-05
source: OpenAlex (CC0)
---

# AudioSet / ESC-50

[Browse dataset profiles](https://scholar4ai.com/datasets/directory.md)

Speech, speaker and sound collections.

- data category: Speech and audio benchmarks
- 0 papers in the last 7 days
- Names searched for: AudioSet, ESC-50

Most-used data: https://scholar4ai.com/rankings/datasets.md · All Speech and audio benchmarks datasets: https://scholar4ai.com/datasets/speech-and-audio-benchmarks.md

## Tasks, access and usage conditions

Mention counts below measure name matches in tracked papers, not benchmark quality or downloads. A family count can cover multiple datasets; it is not a separate count for each comparison row. Choose a task and version first, preserve evaluation splits, and follow the linked release terms. The site's content license does not replace the datasets' own terms.

*Dataset comparison: tasks, access and usage conditions*

| Dataset / version | Task and evaluation | Access and preparation | Usage conditions |
|---|---|---|---|
| AudioSet | Multi-label audio-event detection from short YouTube segments. Evaluation: Use multi-label metrics such as mAP, identify the label set and report recovered clips; do not compare against single-label ESC-50 accuracy. | [Official source / documentation](https://research.google.com/audioset/download.html) Use the provided annotations/features and exact balanced, unbalanced and evaluation splits; source videos may become unavailable. | Google distributes annotations under CC BY 4.0; original YouTube audio/video has separate rights and availability constraints. |
| ESC-50 / ESC-10 | Environmental-sound classification; ESC-50 contains 2,000 five-second clips in 50 classes. Evaluation: Report five-fold accuracy and pretraining; using an arbitrary random clip split can leak source recordings. | [Official source / documentation](https://github.com/karolpiczak/ESC-50) Use the original metadata and five folds, which keep fragments from the same source recording together. | ESC-50 is CC BY-NC; the ESC-10 subset is CC BY. Follow the per-clip attributions in LICENSE. |

Guide updated 2026-11-05.

## Key findings

- From Nov 2023 to Oct 2026, 41 archived papers mentioned AudioSet / ESC-50 in their title, abstract or keywords (the search of earlier months' full texts is still under way), most of them in Jan 2026 (4).
- AudioSet / ESC-50 was mentioned in 1 paper in the last 30 days (2026-10-07 to 2026-11-05), ranking #165 of the 246 datasets mentioned.
- By journal category: Operations Research (1).
- Most of them appeared in Operations Research Forum (1).
- From 2017 to 2026 (so far), AudioSet / ESC-50 was mentioned in 49 papers in the tracked journals, most of them in AI (47), information systems (1) and operations research (1).
- 9 of them were published in 2025, against 0 in 2017.
- Since 2023, the institutions with the most of these papers were Chinese Academy of Sciences (3), Institute of Automation (3) and Shanghai Artificial Intelligence Laboratory (3).
- The journals publishing most of them were International Journal of Computer Vision (9), IEEE Transactions on Pattern Analysis and Machine Intelligence (8) and Expert Systems with Applications (6).
- The most-cited of them is “A comprehensive survey on applications of transformers for deep learning tasks” (Expert Systems with Applications, 2023), with 495 citations.
- Papers use AudioSet / ESC-50 most often with ImageNet (19), YouTube / TikTok (16) and Action recognition (Kinetics, UCF101, HMDB51) (11); it has 8 partners in the code and data network, in the group around ImageNet, CIFAR-10 / CIFAR-100, MNIST / Fashion-MNIST.

## Mentions over 36 months

*Papers mentioning it by publication month, Nov 2023 to Oct 2026 (titles, abstracts and keywords; the full texts of earlier months are still being searched)*

| Trend | Nov 23–Oct 24 | Nov 24–Oct 25 | Nov 25–Oct 26 | 36 months |
|---|---:|---:|---:|---:|
|  | 15 | 10 | 16 | 41 |

## Papers per year

*Papers per publication year and category*

| Category | 2017 | 2018 | 2019 | 2020 | 2021 | 2022 | 2023 | 2024 | 2025 | 2026 (so far) |
|---|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|
| [AI](https://scholar4ai.com/categories/ai.md) | 0 | 0 | 0 | 0 | 1 | 2 | 5 | 18 | 9 | 12 |
| [Information Systems](https://scholar4ai.com/categories/information-systems.md) | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| [Operations Research](https://scholar4ai.com/categories/operations-research.md) | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 |
| All categories | 0 | 0 | 1 | 0 | 1 | 2 | 5 | 18 | 9 | 13 |

## Where the papers come from

*Journals with the most papers, since 2023*

| # | Journal | Papers |
|---:|---|---:|
| 1 | [International Journal of Computer Vision](https://scholar4ai.com/journals/international-journal-of-computer-vision.md) | 9 |
| 2 | [IEEE Transactions on Pattern Analysis and Machine Intelligence](https://scholar4ai.com/journals/ieee-transactions-on-pattern-analysis-and-machine-intelligence.md) | 8 |
| 3 | [Expert Systems with Applications](https://scholar4ai.com/journals/expert-systems-with-applications.md) | 6 |
| 4 | [Neural Computing and Applications](https://scholar4ai.com/journals/neural-computing-and-applications.md) | 5 |
| 5 | [Neural Networks](https://scholar4ai.com/journals/neural-networks.md) | 3 |
| 6 | [Knowledge-Based Systems](https://scholar4ai.com/journals/knowledge-based-systems.md) | 3 |
| 7 | [Information Fusion](https://scholar4ai.com/journals/information-fusion.md) | 3 |
| 8 | [Artificial Intelligence Review](https://scholar4ai.com/journals/artificial-intelligence-review.md) | 2 |
| 9 | [Pattern Recognition](https://scholar4ai.com/journals/pattern-recognition.md) | 1 |
| 10 | [Pattern Recognition Letters](https://scholar4ai.com/journals/pattern-recognition-letters.md) | 1 |

*Institutions with the most papers, since 2023*

| # | Institution | Papers |
|---:|---|---:|
| 1 | [Chinese Academy of Sciences](https://scholar4ai.com/institutions/chinese-academy-of-sciences.md) | 3 |
| 2 | [Institute of Automation](https://scholar4ai.com/institutions/institute-of-automation.md) | 3 |
| 3 | Shanghai Artificial Intelligence Laboratory | 3 |
| 4 | Universitatea Națională de Știință și Tehnologie Politehnica București | 3 |
| 5 | [University of Chinese Academy of Sciences](https://scholar4ai.com/institutions/university-of-chinese-academy-of-sciences.md) | 3 |
| 6 | [University of Oxford](https://scholar4ai.com/institutions/university-of-oxford.md) | 3 |
| 7 | [Harbin Institute of Technology](https://scholar4ai.com/institutions/harbin-institute-of-technology.md) | 2 |
| 8 | [Hefei University of Technology](https://scholar4ai.com/institutions/hefei-university-of-technology.md) | 2 |
| 9 | Institut Polytechnique de Paris | 2 |
| 10 | MIT Computer Science and Artificial Intelligence Laboratory | 2 |

*Countries with the most papers, since 2023*

| # | Country | Papers |
|---:|---|---:|
| 1 | China | 17 |
| 2 | United Kingdom | 7 |
| 3 | United States | 7 |
| 4 | Australia | 4 |
| 5 | Spain | 3 |
| 6 | Romania | 3 |
| 7 | United Arab Emirates | 2 |
| 8 | Finland | 2 |
| 9 | France | 2 |
| 10 | Hong Kong | 2 |

*Keywords with the most papers, since 2023*

| # | Keyword | Papers |
|---:|---|---:|
| 1 | ESC-50 | 5 |
| 2 | [self-attention](https://scholar4ai.com/keywords/self-attention.md) | 5 |
| 3 | audio classification | 4 |
| 4 | AudioSet | 3 |
| 5 | audiovisual integration | 3 |
| 6 | [convolutional neural networks](https://scholar4ai.com/keywords/convolutional-neural-networks.md) | 3 |
| 7 | cross-modal alignment | 3 |
| 8 | [multimodal fusion](https://scholar4ai.com/keywords/multimodal-fusion.md) | 3 |
| 9 | multimodal learning | 3 |
| 10 | temporal modeling | 3 |

## Most-cited papers

*Most-cited papers since 2017*

| # | Citations | Per month | Title | Journal | Published | Authors |
|---:|---:|---:|---|---|---|---|
| 1 | 495 | 14.0 | [A comprehensive survey on applications of transformers for deep learning tasks](https://doi.org/10.1016/j.eswa.2023.122666) | [Expert Systems with Applications](https://scholar4ai.com/journals/expert-systems-with-applications.md) | 2023-11-22 | Saidul Islam, Hanae Elmekki, Ahmed Elsebai et al. |
| 2 | 212 | 6.9 | [Automatic speech recognition using advanced deep learning approaches: A survey](https://doi.org/10.1016/j.inffus.2024.102422) | [Information Fusion](https://scholar4ai.com/journals/information-fusion.md) | 2024-04-15 | Hamza Kheddar, Mustapha Hemis, Yassine Himeur |
| 3 | 102 | 4.3 | [Advances in ECG and PCG-based cardiovascular disease classification: a review of deep learning and machine learning methods](https://doi.org/10.1186/s40537-024-01011-7) | [Journal Of Big Data](https://scholar4ai.com/journals/journal-of-big-data.md) | 2024-11-07 | Asmaa Ameen, Ibrahim Eldesouky Fattoh, Tarek Abd El‐Hafeez et al. |
| 4 | 81 | 1.6 | [Environmental Sound Classification on the Edge: A Pipeline for Deep Acoustic Networks on Extremely Resource-Constrained Devices](https://doi.org/10.1016/j.patcog.2022.109025) | [Pattern Recognition](https://scholar4ai.com/journals/pattern-recognition.md) | 2022-09-07 | Md Mohaimenuzzaman, Christoph Bergmeir, Ian Thomas West et al. |
| 5 | 74 | 3.0 | [Audio-Visual Segmentation with Semantics](https://doi.org/10.1007/s11263-024-02261-x) | [International Journal of Computer Vision](https://scholar4ai.com/journals/international-journal-of-computer-vision.md) | 2024-10-14 | Jinxing Zhou, Xuyang Shen, Jianyuan Wang et al. |

## Used together with

Papers use AudioSet / ESC-50 most often with ImageNet (19), YouTube / TikTok (16) and Action recognition (Kinetics, UCF101, HMDB51) (11); it has 8 partners in the code and data network, in the group around ImageNet, CIFAR-10 / CIFAR-100, MNIST / Fashion-MNIST.

*What papers use AudioSet / ESC-50 together with (at least 3 papers; lift 1 = chance)*

| Used with | Type | Papers using both | Lift |
|---|---|---:|---:|
| [ImageNet](https://scholar4ai.com/datasets/imagenet.md) | Dataset | 19 | 7.9 |
| [YouTube / TikTok](https://scholar4ai.com/datasets/youtube-tiktok.md) | Dataset | 16 | 8.5 |
| [Action recognition (Kinetics, UCF101, HMDB51)](https://scholar4ai.com/datasets/action-recognition-kinetics-ucf101-hmdb51.md) | Dataset | 11 | 37.0 |
| [LibriSpeech](https://scholar4ai.com/datasets/librispeech.md) | Dataset | 5 | 80.0 |
| [MS COCO](https://scholar4ai.com/datasets/ms-coco.md) | Dataset | 5 | 5.3 |
| [CIFAR-10 / CIFAR-100](https://scholar4ai.com/datasets/cifar-10-cifar-100.md) | Dataset | 3 | 2.1 |
| [PASCAL VOC](https://scholar4ai.com/datasets/pascal-voc.md) | Dataset | 3 | 8.8 |
| [UCI Machine Learning Repository](https://scholar4ai.com/datasets/uci-machine-learning-repository.md) | Dataset | 3 | 4.7 |

The whole network: https://scholar4ai.com/code/network.md

## Papers

*Papers mentioning AudioSet / ESC-50 in the last 30 days, newest first.*

| Date | Title | Journal | Authors | Affiliations |
|---|---|---|---|---|
| 2026-10-15 | [Transformer Architectures for Respiratory Sound Analysis and Multimodal Diagnosis](https://doi.org/10.1007/s43069-026-00698-z) | [Operations Research Forum](https://scholar4ai.com/journals/operations-research-forum.md) | Theodore Aptekarev, Vladimir Sokolovsky, G. B. Furman et al. | National Research University Higher School of Economics; Ben-Gurion University of the Negev; Perm State Medical Academy |

## Cite this page

Scholar4AI. (2026). AudioSet / ESC-50. Journal and Conference Research Rankings. Retrieved November 5, 2026, from https://scholar4ai.com/datasets/audioset-esc-50

```bibtex
@misc{scholar4ai-datasets-audioset-esc-50,
  author = {{Scholar4AI}},
  title = {AudioSet / ESC-50},
  year = {2026},
  howpublished = {\url{https://scholar4ai.com/datasets/audioset-esc-50}},
  urldate = {2026-11-05},
  note = {Journal and Conference Research Rankings; data from OpenAlex (CC0); licensed CC BY-NC 4.0}
}
```

License: CC BY-NC 4.0 (https://creativecommons.org/licenses/by-nc/4.0/).