Datasets:

Modalities:
Text
Formats:
csv
ArXiv:
Libraries:
Datasets
pandas
License:
sylviali commited on
Commit
15793b3
1 Parent(s): d2f3374

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +6 -6
README.md CHANGED
@@ -5,20 +5,20 @@ license: mit
5
  # EDEN ASR Dataset
6
  A subset of this data was used to support the development of empathetic feedback modules in [EDEN](https://arxiv.org/abs/2406.17982) and [its prior work](https://arxiv.org/abs/2404.13764).
7
 
8
- To create data for testing various modules, we utilized audio clips of native Mandarin speakers conversing with a chatbot collected from an [English practice platform](https://dl.acm.org/doi/abs/10.1145/3491140.3528329?casa_token=ER-mfy0xauQAAAAA:FyDgmH0Y0ke7a6jpOnuycP1HRfeV1B5qaq5JWM5OV5dB9fLFL_vzVRUacZ4fUMRBDl71UeWMIA9Z).
9
  3081 audio clips from 613 conversations and 163 users remained after filtering.
10
  The filtering process removes audio clips containing only Mandarin, duplicates, and a subset of self-introductions from the users.
11
- Each audio clip ranges from one second to two minutes. We did not collect demographic information for user identity protection.
12
 
13
- In our original work, we directly transcribed the speech with Whisper Medium. However, since the audio clips are accented speech, there exist instances of ASR error in these transcripts. We have **manually verified** the original transcripts to ensure they are high-quality.
14
 
15
  ## Dataset Columns
16
- * **audio_url**: The URL to the audio clips; you can download them using wget to your local machines.
17
  * **emotion_label**: We manually labeled a subset of the clips as **Neutral** (neutral emotions), **Negative** (the speaker is displaying negative emotions), or **Pauses** (there are a lot of pauses in the speech, potentially signaling language anxiety). The process is documented in [EDEN's prior work](https://arxiv.org/abs/2404.13764). When this field is empty, it means that the clip was not labeled.
18
  * **corrected_whisper_transcript**: The high-quality transcript we have verified; the person performing verification is a native English speaker.
19
 
20
  ## Intended Use
21
- This set of data can be used for Accented ASR research!
22
 
23
  ## Code Example
24
 
@@ -38,7 +38,7 @@ print(dataset[0]["corrected_whisper_transcript"])
38
  # Extract the emotion label
39
  print(dataset[0]["emotion_label"])
40
 
41
- # Check negative emotion audio clips
42
  negative_emotion_clips = dataset.filter(lambda example: example["emotion_label"] == "Negative")
43
  print(len(negative_emotion_clips))
44
  print(negative_emotion_clips[0])
 
5
  # EDEN ASR Dataset
6
  A subset of this data was used to support the development of empathetic feedback modules in [EDEN](https://arxiv.org/abs/2406.17982) and [its prior work](https://arxiv.org/abs/2404.13764).
7
 
8
+ The dataset contains audio clips of native Mandarin speakers. The speakers conversed with a chatbot hosted on an [English practice platform](https://dl.acm.org/doi/abs/10.1145/3491140.3528329?casa_token=ER-mfy0xauQAAAAA:FyDgmH0Y0ke7a6jpOnuycP1HRfeV1B5qaq5JWM5OV5dB9fLFL_vzVRUacZ4fUMRBDl71UeWMIA9Z).
9
  3081 audio clips from 613 conversations and 163 users remained after filtering.
10
  The filtering process removes audio clips containing only Mandarin, duplicates, and a subset of self-introductions from the users.
11
+ Each audio clip ranges from one second to two minutes. We did not collect demographic information to protect user identities.
12
 
13
+ In our original work, we directly transcribed the speech with Whisper Medium. However, since the audio clips are accented speech, these transcripts have instances of ASR error. We have **manually verified** the original transcripts to ensure they are high-quality.
14
 
15
  ## Dataset Columns
16
+ * **audio_url**: The URL to the audio clips; you can download them using wget or urllib to your local machines (see the code snippet below).
17
  * **emotion_label**: We manually labeled a subset of the clips as **Neutral** (neutral emotions), **Negative** (the speaker is displaying negative emotions), or **Pauses** (there are a lot of pauses in the speech, potentially signaling language anxiety). The process is documented in [EDEN's prior work](https://arxiv.org/abs/2404.13764). When this field is empty, it means that the clip was not labeled.
18
  * **corrected_whisper_transcript**: The high-quality transcript we have verified; the person performing verification is a native English speaker.
19
 
20
  ## Intended Use
21
+ Accented ASR research!
22
 
23
  ## Code Example
24
 
 
38
  # Extract the emotion label
39
  print(dataset[0]["emotion_label"])
40
 
41
+ # Check audio clips labeled as having negative emotions
42
  negative_emotion_clips = dataset.filter(lambda example: example["emotion_label"] == "Negative")
43
  print(len(negative_emotion_clips))
44
  print(negative_emotion_clips[0])