Score Task/Task Achievement

Run in postman: Score Task Achievement

The Speechace Task Achievement API supports following task types:

• Describe-Image: The speaker is presented with an image and asked to describe the details, relationships, and conclusion to be drawn from elements of the image.

• Retell-Lecture: The speaker listens to a 1-2 minute lecture and is asked to summarize the lecture focusing on key elements, concepts and conclusions from the lecture.

• Answer-Question: The speaker is presented with a short question which typically requires a one or two word answer.

Each task type has particular input and outputs:

Task Type	Inputs	Outputs
describe-image	task_context: A model description of the image which is presented to the speaker. Max length: 1024 chars.	Task score on scale of 0-5.
retell-lecture	task_context: A model summary of the lecture which is presented to the speaker. Max length: 1024 chars.	Task score on scale of 0-5.
answer-question	task-question: The question presented to the user.	Task score on scale of 0-1 where 0 is incorrect and 1 is correct.

The API supports different modes in combining task scores and language scores in assessment:

1. user_audio_file or user_audio_text: The speaker's response can be submitted as either audio or text, allowing task scoring to be used with written responses as well.

2. include_speech_score: Speech scoring can be included or excluded along with the task score. Note that if user_audio_text is used, the include_speech_score will always be zero. Therefore, in written responses, only task scores are provided.

All tasks are available in the following languages:

• English (en-us, en-gb)

• Spanish (es-es, es-mx)

• French (fr-fr, fr-ca)

Request Format

The endpoint that is to be used will depend on the region of your subscription. For example, for US West, the endpoint is https://api.speechace.co.

POST https://api.speechace.co/api/scoring/task/v9/json

Example with audio

curl --location -g 'https://api.speechace.co/api/scoring/task/v9/json?key={{speechace_premiumkey}}&task_type=describe-image&dialect=en-us' \
--form 'task_context="This bar chart illustrates the declining trend related to the percent of U.S. workforce engaged in farm labor in the 19th century. In 1840, for example, around 69% of the U.S. workforce was engaged in farm labor; in 1860, almost 60% of the U.S. workforce was engaged in farm labor; in 1880, only 50% of the U.S. workforce was engaged in farm labor; and in 1900, less than 40% of the U.S. workforce was engaged in farm labor."' \
--form 'user_audio_file=@"barchataudiofile.mp3"' \
--form 'include_speech_score="1"'

Example with text

Query Parameters

Parameter	Type	Description
key	String	API key issued by Speechace.
dialect	String	This is the dialect in which the speaker will be assessed. Supported values are: en-us, en-gb, fr-fr, fr-ca, es-es, es-mx.
user_id	String	Optional: A unique anonymized identifier (generated by your applications) for the end-user who spoke the audio.
task_type	String	The task_type to score. Supported types are: describe-image retell-lecture answer-question.

Request Body

Parameter	Type	Description
task_context	String	The context or model or model answer for the task presented to the speaker. Used in the following task-types: describe-image: a model description of the image retell-lecture: a model description of the lecture This must be provided in the same language as the one being assessed.
task_question	String	The task question presented to the speaker, used in task-type = answer-question. This must be provided in the same language as the one being assessed.
user_audio_file	File	file with user audio (wav, mp3, m4a, webm, ogg, aiff)
include_speech_score	String	Set to 1, to include scoring other aspects of the speech: Pronunciation, Fluency, Grammar, Vocab, Coherence. Set to 0 if you only want to receive the task score only.
user_audio_text	String	A text transcript of the speaker's response. Use this field instead of user_audio_file if you already have a transcript of the user's response and do not wish to re-transcribe an audio. Note: In this case, you will only be able to receive an overall task_score.

Response Example

Notice the task_score.score key for the overall task achievement score in the response below:

Task = describe-image

{
  "status": "success",
  "task_score": {
    "type": "describe-image",
    "version": "0.1",
    "score": 4,
    "transcript": "This bar graph shows the percent of US workforce engaged in farm labor, and that's data from 1840 to 1900. Ear now starting with 1840, the percentage was 70 percentage. After that there is a gradual decrease in the number of workforce engaged in farm labor to 60 percentage in 1860 and further down to 18 around 50% in 1880. And then in 1900 it decreased to 40 percentage. Overall, there is a continuous decrease in the engagement in the farm sector."
  },
  "quota_remaining": -1,
  "speech_score": {
    "transcript": "This bar graph shows the percent of US workforce engaged in farm labor, and that's data from 1840 to 1900. Ear now starting with 1840, the percentage was 70 percentage. After that there is a gradual decrease in the number of workforce engaged in farm labor to 60 percentage in 1860 and further down to 18 around 50% in 1880. And then in 1900 it decreased to 40 percentage. Overall, there is a continuous decrease in the engagement in the farm sector.",
    "word_score_list": [<.....pronunciation metrics>],
    "ielts_score": {<....ielts scores>},
    "pte_score": {<...pte scores>},
    "speechace_score": {<...speechace scores>},
    "toeic_score": {<...toeic scores>},
    "cefr_score": {
      "pronunciation": "B2",
      "fluency": "B2",
      "grammar": "B1+",
      "coherence": "B1",
      "vocab": "B1+",
      "overall": "B1+"
    },
    "fluency": {<...fluency metrics>},
    "asr_version": "0.4"
  },
  "version": "9.7"
}

Task = retell-lecture

Task = answer-question

The pronunciation and fluency interpretation of the key elements in the response of the spoken word or sentence remains the same.

The new addition is the task score parameters, which indicate the extent to which the task has been achieved.

Difference between task_context and relevance_context

• Relevance is binary and is higher level. It evaluates whether the response is on-topic or not (True or False)

• Task Achievement is more nuanced and scores how well the response addresses the task

For a general question such as "Do you think the government should subsidize healthcare?" relevance is primarily assessed, as there is no definitive right or wrong answer; the focus is on whether the response is on topic.

In contrast, for a specific question like "What does the following business chart tell us?" a specific answer is expected. Therefore, a nuanced task context and detailed task score are required to evaluate how well the response addresses the specific elements of the task.