Answer span extraction in NLP - Model Pipeline Trace

Model Pipeline - Answer span extraction

This pipeline finds the exact part of a text that answers a question. It reads the question and text, then predicts the start and end positions of the answer inside the text.

Data Flow - 5 Stages

1Input data

1000 samples x 2 texts (question and context)→Raw question and context pairs→1000 samples x 2 texts

Question: 'Where is the Eiffel Tower?'; Context: 'The Eiffel Tower is in Paris, France.'

↓

2Tokenization

1000 samples x 2 texts→Split texts into tokens and convert to numbers→1000 samples x 128 tokens

Tokens: ['[CLS]', 'Where', 'is', 'the', 'Eiffel', 'Tower', '?', '[SEP]', 'The', 'Eiffel', 'Tower', 'is', 'in', 'Paris', ',', 'France', '.', '[SEP]']

↓

3Model input preparation

1000 samples x 128 tokens→Create input IDs, attention masks, and token type IDs→1000 samples x 128 tokens x 3 arrays

Input IDs: [101, 2073, 2003, 1996, 3000, 2433, 1029, 102, 1996, 3000, 2433, 2003, 1999, 3000, 1010, 3000, 1012, 102]

↓

4Model prediction

1000 samples x 128 tokens x 3 arrays→Predict start and end logits for answer span→1000 samples x 128 start logits + 128 end logits

Start logits: [0.1, 0.2, ..., 5.0, ..., 0.1]; End logits: [0.1, 0.1, ..., 4.8, ..., 0.2]

↓

5Answer span extraction

1000 samples x 128 start logits + 128 end logits→Select tokens with highest start and end logits to form answer→1000 samples x answer text

Answer: 'Paris, France'

Training Trace - Epoch by Epoch

Loss
1.2 |*       
0.8 |  *     
0.5 |    *   
0.3 |      * 
0.25|       *
    +--------
     1 2 3 4 5 Epochs

Epoch	Loss ↓	Accuracy ↑	Observation
1	1.2	0.45	Model starts learning, loss high, accuracy low
2	0.8	0.60	Loss decreases, accuracy improves
3	0.5	0.75	Model learns better answer spans
4	0.3	0.85	Good convergence, loss low, accuracy high
5	0.25	0.88	Training stabilizes with good performance

Prediction Trace - 4 Layers

Layer 1: Tokenization

Layer 2: Model input preparation

Layer 3: Model prediction

Layer 4: Answer span extraction

Model Quiz - 3 Questions

What does the model predict to find the answer in the text?

A The full text of the answer directly

B Start and end positions of the answer span

C Only the start position of the answer

D The question rewritten

Key Insight

Answer span extraction models learn to find exact start and end points of answers in text by predicting positions, not by generating text. This makes them precise for question answering tasks.

Practice

1. What is the main goal of answer span extraction in NLP?

easy

A. To generate new text based on a prompt

B. To find the exact part of text that answers a question

C. To summarize long documents into short sentences

D. To translate text from one language to another

Solution

1. Step 1: Understand the purpose of answer span extraction

   Answer span extraction focuses on locating the exact segment in a text that directly answers a question.

2. Step 2: Compare with other NLP tasks

   Unlike translation, summarization, or text generation, answer span extraction pinpoints a specific text span as the answer.

3. Final Answer:

   To find the exact part of text that answers a question -> Option B

4. Quick Check:

   Answer span extraction = find exact answer span [OK]

Hint: Answer span extraction locates exact text answers [OK]

Common Mistakes:

• Confusing answer span extraction with translation
• Thinking it summarizes text instead of extracting spans
• Assuming it generates new text

2. Which of the following is the correct way to represent the start and end positions for answer span extraction in code?

easy

A. start_index and end_index as integers

B. start_word and end_word as strings

C. start_time and end_time as floats

D. start_char and end_char as booleans

Solution

1. Step 1: Identify typical data types for positions

   Positions in text are usually represented by integer indices marking start and end locations.

2. Step 2: Evaluate options

   Strings or booleans do not represent positions well; floats for time are unrelated to text spans.

3. Final Answer:

   start_index and end_index as integers -> Option A

4. Quick Check:

   Positions = integer indices [OK]

Hint: Positions in text are integer indices [OK]

Common Mistakes:

• Using strings instead of integer indices
• Confusing character positions with time values
• Using booleans for position markers

3. Given the text: 'The cat sat on the mat.' and predicted start index = 1, end index = 4, what is the extracted answer span?

medium

A. 'cat sat on'

B. 'sat on the'

C. 'on the mat'

D. 'The cat sat'

Solution

1. Step 1: Identify tokens and their indices

   Tokenizing the sentence: ['The'(0), 'cat'(1), 'sat'(2), 'on'(3), 'the'(4), 'mat.'(5)]. The indices given (1 to 4) refer to 0-based token positions.

2. Step 2: Extract tokens from start to end index

   In standard extraction, take tokens[start:end] (end exclusive): tokens[1:4] = ['cat'(1), 'sat'(2), 'on'(3)] = 'cat sat on'.

3. Final Answer:

   'cat sat on' -> Option A

4. Quick Check:

   Extract tokens from start to end index = 'cat sat on' [OK]

Hint: Match indices to tokens carefully [OK]

Common Mistakes:

• Confusing character indices with token indices
• Off-by-one errors in slicing
• Ignoring punctuation in tokens

4. You have a model that predicts start and end indices for answer spans but sometimes the end index is smaller than the start index. What is the best way to fix this bug?

medium

A. Ignore the prediction and return an empty answer

B. Always set end index to start index plus one

C. Swap the start and end indices if end < start

D. Use only the start index as the answer

Solution

1. Step 1: Understand the problem with indices

   End index smaller than start index is invalid because answer spans must go forward in text.

2. Step 2: Choose a fix that preserves valid spans

   Swapping start and end indices corrects the order and keeps the predicted span meaningful.

3. Final Answer:

   Swap the start and end indices if end < start -> Option C

4. Quick Check:

   Fix invalid spans by swapping indices [OK]

Hint: Swap indices if end < start to fix spans [OK]

Common Mistakes:

• Ignoring invalid spans instead of fixing
• Forcing fixed span length blindly
• Using only one index loses answer context

5. In a question-answering system, the model outputs start logits and end logits for each token. How should you combine these to find the best answer span?

hard

A. Choose random start and end indices

B. Pick the token with the highest start logit only

C. Pick the token with the highest end logit only

D. Find the pair of start and end indices with the highest sum of start and end logits where start ≤ end

Solution

1. Step 1: Understand logits for start and end tokens

   Start and end logits represent scores for each token being the start or end of the answer span.

2. Step 2: Combine logits to find best span

   We look for the pair (start, end) with the highest combined score, ensuring start ≤ end to form a valid span.

3. Final Answer:

   Find the pair of start and end indices with the highest sum of start and end logits where start ≤ end -> Option D

4. Quick Check:

   Combine start and end logits to find best span [OK]

Hint: Sum start and end logits, ensure start ≤ end [OK]

Common Mistakes:

• Ignoring end logits and using start only
• Choosing invalid spans where end < start
• Picking random indices without scores