audio
audioduration (s)
1.5
8.19
intent_class
int64
0
13
template
stringclasses
164 values
speaker_id
int64
1
11
6
Please tell me the IFSC Code
3
7
Credit card pin generation
11
3
Is my card dispatched?
5
11
I want to change the limit of my online transactions on my (credit/debit) card
11
8
Amount deducted from account without my knowledge
3
0
Where is the bank branch ?
3
13
I want to report my lost debit card
2
3
Is my card dispatched?
11
4
Give me the outstanding balance on my credit card
2
7
Debit card pin generation
11
10
I want to know my balance
11
3
I want to know my credit card dispatch status
8
13
My card is lost
1
8
Amount deducted from account without my knowledge
11
12
I want to block my UPI
1
9
I want to get loan
5
2
Last <numeric> transaction details
3
9
Home loan
8
7
How to generate pin for credit card
2
6
How to know my IFSC Code
3
3
When will you deliver my credit card
11
7
Debit card pin generation
6
2
Please give me the details of my last transaction
2
9
I want to speak to loan department
10
2
What was the amount of my last transaction
8
7
how to generate credit card pin
9
7
How to generate pin for credit card
6
4
Give me the outstanding balance on my credit card
4
8
I want to report a fraud
4
4
what is the outstanding balance on my credit card?
2
11
How to change the limit of credit card
8
4
What is my pending amount on my credit card bill
6
11
I want to change the limit of my e-statement on my (credit/debit) card
2
0
Please tell me the nearest branch of the bank
10
10
I want my account balance
1
8
I want to report a fraud
1
13
I lost my credit card
6
3
I want to know my credit card dispatch status
1
6
IFSC Code
6
4
Total outstanding balance?
2
8
Amount deducted from account without my knowledge
3
8
Amount deducted from account without my knowledge
5
1
activate debit card
4
12
I want to block mobile banking
6
5
I am unable to use my credit card
4
11
How to change the limit of credit card
1
2
Past transaction details
11
6
I want my IFSC code
3
5
There is a problem with my credit card
2
7
generate credit card pin
5
7
I want to set my credit card pin
2
6
I want to know my IFSC Code
6
8
I want to report a wrong transaction
2
9
How to get personal loan
4
1
How to activate credit card
5
9
I want to know about home loan
5
13
My card is lost
5
12
I wanted to block my mobile banking
2
0
What is the nearest branch?
2
0
Where is the bank branch ?
5
1
activate
2
1
I want to activate my debit card
4
0
Where is the branch located?
8
13
Credit card is missing
1
10
I want to know my balance
10
10
I want to know my balance
4
8
Fraud transaction happened from my account
10
7
I want to generate pin for my credit card
8
8
I want to report a wrong transaction
10
12
Block credit card
4
10
I want to know my balance
1
8
Unauthorised transaction
3
2
I want to know about my last <numeric> transactions
1
2
Details of last transaction
7
2
details of last few transactions
5
12
I wanted to block my mobile banking
4
4
Give me the outstanding balance on my credit card
6
3
when will my card be delivered?
11
1
I want to start my credit card
2
2
details of last few transactions
6
10
balance enquiry
5
11
I want to change the limit of my (credit/debit) card
4
0
Where is the bank branch ?
4
9
I want to know about personal loan
11
5
Debit card issue
7
6
IFSC Code
7
7
I want to set my credit card pin
5
5
I am unable to pay with my credit card
8
12
I want to block my UPI
3
10
bank account balance enquiry
2
9
I have a uery regarding loan
6
7
Debit card pin generation
7
4
How to know my credit card outstanding balance
2
2
Details of last transaction
5
6
What is my IFSC Code
8
2
I want to know about my last <numeric> transactions
10
5
There is a problem with my credit card
6
12
Can I block my credit card
5
5
My debit card is not working
5
4
What is my pending amount on my credit card bill
1

Skit-S2I is a Speech to Intent dataset for Indian English (en-IN), that covers 14 coarse-grained intents from the Banking domain. This work is inspired by a similar release in the Minds-14 dataset - here, we restrict ourselves to Indian English but with a larger training set. The dataset is split into:

  • test - 100 samples per intent
  • train - >650 samples per intent

The data was generated by 11 Indian speakers, recording over a telephony line. We also provide access to anonymised speaker information - like gender, languages spoken, native language - to enable more structured discussions around robustness and bias, in the models you train.

This Datasheet follows from the Datasheets for datasets paper.

Motivation

Q1) For what purpose was the dataset created ? Was there a specific task in mind ? Was there a specific gap that needed to be filled ?

Ans. This is a dataset for Intent classification from (Indian English) speech, and covers 14 coarse-grained intents from the Banking domain. While there are other datasets that have approached this task, here we provide a much largee training dataset (>650 samples per intent) to train models in an end-to-end fashion. We also provide anonymised speaker information to help answer questions around model robustness and bias.

Q2) Who created the dataset and on behalf of which entity ?

Ans. The (internal) Operations team at Skit was involved in the generation of the dataset, and provided their information for (anonymous) release. Unnati Senani was involved in the curation of utterance templates, and Kriti Anandan and Kumarmanas Nethil were involved in the planning and collection of utterances - using an internal tool called sandbox. These contributors worked on this dataset as part of the Conversational UX and ML teams at Skit.

Q3) Who funded the creation of the dataset ?

Ans. Skit funded the creation of this dataset.

Composition

Q4) What do the instances that comprise the dataset consist of ?

Ans. The intent dataset is split across train.csv and test.csv. In both, individual instances consist of the following fields:

  • id
  • intent_class
  • template
  • audio_path
  • speaker_id

You can trace more information on the intents, using the shared intent_class field in intent_info.csv:

  • intent_class
  • intent_name
  • description

You can trace more information on the speakers, using the shared speaker_id field in speaker_info.csv:

  • speaker_id
  • native_language
  • languages_spoken
  • places_lived
  • gender

Q5) How many instances are there in total (of each type, if appropriate) ?

Ans. In all there are 11845 samples, across the train and test splits:

  • test.csv has a total of 1400 samples, with exactly 100 samples per intent
  • train.csv has a total of 10445 samples, with atleast 650 samples per intent

The 11 speakers are distributed across the dataset, but unequally. However:

  • each intent has data from all speakers
  • the speakers are stratified across the train and test split - for each intent independently

Some statistics on the speakers are provided below. More granular information can be found in speaker_info.csv:

  • Native languages: Hindi(4), Bengali(3), Kannada(2), Malayalam(1), Punjabi(1)
  • Languages spoken: Hindi, English, Bengali, Odia, Kannada, Punjabi, Malayalam, Bihari, Marathi
  • Indian states lived in: Bihar, Odisha, Karnataka, West Bengal, Punjab, Kerala, Jharkhand, Maharashtra

Q6) Does the dataset contain all possible instances or is it a sample (not necessarily random) of instances from a larger set ?

Ans. For each intent, our Conversational UX team generated a list of templates. These are meant to be a (satisfactory) representation of all the variations in utterances, seen in human speech. These templates were used as a guide by the speakers when generating data. So, this dataset is limited by the templates and the variations that speakers added (spontaneously).

Q7) Are there recommended data splits (e.g., training, development/validation, testing) ?

Ans. The recommended split into train and test sets is provided as train.csv and test.csv respectively.

Q8) Are there any errors, sources of noise, or redundancies in the dataset?

Ans. There could be channel noise present in the dataset, because the data was generated through telephone calls. However, background noise will not be as prevalent as in real-world scenarios, since these telephone calls were made in a semi-controlled environment.

Q9) Other comments.

Ans. Speakers were responsible for generating variations in utterances, using the template field as a guide. So, there could be some unintentional overlap across the content of utterances.

Collection Process

Q10) How was the data associated with each instance acquired ?

Ans. Members of the (internal) Operation team generated each utterance - using the associated template field as a guide, and injecting their own variations into the speech utterance.

Q11) Who was involved in the data collection process and how were they compensated ?

Ans. The data was generated by the (internal) Operations team and they are/were full-time employees.

Q12) Over what timeframe was the data collected ?

Ans. This data was collected over a time period of 1 month.

Q13) Was any preprocessing/cleaning/labelling of the data done ?

Ans. Audio instances in the dataset were auto-labelled with their associated intent and template fields. For more information on this, refer to the documentation of sandbox.

Recommended Uses

Q14) Has the dataset been used for any tasks already ?

Ans. It has been used to benchmark models for the task of intent classification from speech.

Q15) What (other) tasks could the dataset be used for ?

Ans. We provide speaker characteristics. So, this dataset could be used for alternate classification tasks from speech - like, gender or native language.

Distribution and Maintenance

Q16) Will the dataset be distributed under a copyright or other intellectual property (IP) license ?

Ans. This dataset is being distributed under a CC BY NC license.

Q17) Who will be maintaining the dataset ?

Ans. The research team at Skit will be maintaining the dataset. They can be contacted by sending an email to [email protected].

Q18) Will the dataset be updated in the future (e.g., to correct labelling errors, add new instances, delete instances) ?

Ans. Incase there are errors, we will try to collate and share an updated version every 3 months. We also plan to add more instances and variations to the dataset - to make it more robust.

Downloads last month
64
Edit dataset card