Open-domain QA basics in NLP - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to load a pre-trained question answering model using Hugging Face Transformers.

NLP

from transformers import pipeline
qa_pipeline = pipeline('[1]')

Drag options to blanks, or click blank then click option'

Aquestion-answering

Btext-generation

Csentiment-analysis

Dtranslation

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'text-generation' instead of 'question-answering'.

Confusing sentiment analysis with question answering.

The pipeline type for open-domain question answering is 'question-answering'.

2fill in blank

medium

Complete the code to get the answer from the QA pipeline given a question and context.

NLP

result = qa_pipeline({'question': '[1]', 'context': 'The Eiffel Tower is in Paris.'})

Drag options to blanks, or click blank then click option'

AWhen was the Eiffel Tower built?

BWhat is the capital of France?

CWhere is the Eiffel Tower?

DWho built the Eiffel Tower?

Attempts:

3 left

💡 Hint

Common Mistakes

Asking about the builder or date which is not in the context.

Using a question unrelated to the context.

The question 'Where is the Eiffel Tower?' matches the context about the Eiffel Tower's location.

3fill in blank

hard

Fix the error in the code to extract the answer text from the result dictionary.

NLP

answer_text = result['[1]']

Drag options to blanks, or click blank then click option'

Atext

Banswer

Cresponse

Dresult

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'text' or 'response' which are not keys in the result.

Trying to access 'result' key inside the result dictionary.

The key 'answer' contains the predicted answer text in the result dictionary.

4fill in blank

hard

Fill both blanks to create a dictionary with question and context for the QA pipeline.

NLP

input_data = {'[1]': 'What is AI?', '[2]': 'AI stands for Artificial Intelligence.'}

Drag options to blanks, or click blank then click option'

Aquestion

Banswer

Ccontext

Dresponse

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'answer' or 'response' as keys instead of 'question' and 'context'.

Mixing up the order of keys.

The QA pipeline expects keys 'question' and 'context' in the input dictionary.

5fill in blank

hard

Fill all three blanks to run the QA pipeline and print the answer.

NLP

result = qa_pipeline({'[1]': '[2]', '[3]': 'Python is a programming language.'})
print(result['answer'])

Drag options to blanks, or click blank then click option'

Aquestion

BWhat is Python?

Ccontext

Danswer

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'answer' as a key in the input dictionary.

Putting the question in the 'context' key.

The input dictionary keys are 'question' and 'context', and the question is 'What is Python?'.