Attention mechanism basics in NLP

Introduction

Attention helps a model focus on important parts of input when making decisions. It works like how we pay attention to key words in a sentence to understand its meaning.

Translating a sentence from one language to another

Answering questions based on a paragraph of text

Summarizing a long article into a short summary

Recognizing objects in an image by focusing on important areas

Syntax

NLP

attention_scores = query @ key.T / sqrt(d_k)
attention_weights = softmax(attention_scores)
output = attention_weights @ value

query, key, and value are vectors or matrices representing parts of the input.

The division by sqrt(d_k) helps keep the scores stable.

Examples

This example shows how to calculate attention weights and output using simple vectors.

NLP

import torch
import torch.nn.functional as F

query = torch.tensor([[1., 0., 1.]])
key = torch.tensor([[1., 0., 0.], [0., 1., 0.]])
value = torch.tensor([[1., 2.], [3., 4.]])

scores = query @ key.T / (3 ** 0.5)
weights = F.softmax(scores, dim=1)
output = weights @ value
print(output)

This example uses numpy to do the same attention calculation without PyTorch.

NLP

import numpy as np

def softmax(x):
    e_x = np.exp(x - np.max(x))
    return e_x / e_x.sum(axis=-1, keepdims=True)

query = np.array([1, 0, 1])
key = np.array([[1, 0, 0], [0, 1, 0]])
value = np.array([[1, 2], [3, 4]])

scores = query @ key.T / np.sqrt(3)
weights = softmax(scores)
output = weights @ value
print(output)

Sample Model

This program shows a simple attention mechanism calculation step-by-step using PyTorch. It prints the scores, weights, and final output vector.

NLP

import torch
import torch.nn.functional as F

# Define query, key, value tensors
query = torch.tensor([[1., 0., 1.]])  # shape (1, 3)
key = torch.tensor([[1., 0., 0.], [0., 1., 0.]])  # shape (2, 3)
value = torch.tensor([[1., 2.], [3., 4.]])  # shape (2, 2)

d_k = query.size(-1)  # dimension of key vectors

# Calculate attention scores
scores = query @ key.T / (d_k ** 0.5)  # shape (1, 2)

# Apply softmax to get attention weights
weights = F.softmax(scores, dim=1)  # shape (1, 2)

# Multiply weights by values to get output
output = weights @ value  # shape (1, 2)

print(f"Attention scores: {scores}")
print(f"Attention weights: {weights}")
print(f"Output: {output}")

Important Notes

Attention helps models decide what to focus on, improving understanding.

Softmax turns scores into probabilities that add up to 1.

Query, key, and value come from the input data or previous layers.

Summary

Attention finds important parts of input to focus on.

It uses query, key, and value vectors to calculate weighted outputs.

Softmax makes scores into weights that sum to one.

Practice

1. What is the main purpose of the attention mechanism in NLP models?

easy

A. To reduce the number of layers in the model

B. To focus on important parts of the input data

C. To increase the size of the input data

D. To randomly shuffle the input tokens

Solution

1. Step 1: Understand the role of attention

   Attention helps the model decide which parts of the input are important to look at when making predictions.

2. Step 2: Compare options with the concept

   Only To focus on important parts of the input data correctly describes this focus on important input parts.

3. Final Answer:

   To focus on important parts of the input data -> Option B

4. Quick Check:

   Attention = Focus on important input [OK]

Hint: Attention means focusing on key input parts [OK]

Common Mistakes:

• Thinking attention increases input size
• Confusing attention with model depth
• Assuming attention shuffles data

2. Which of the following correctly represents the formula to compute attention weights using query (Q) and key (K) vectors?

easy

A. Sigmoid(Q - K)

B. Softmax(Q + K)

C. ReLU(Q x K)

D. Softmax(Q x K^T)

Solution

1. Step 1: Recall attention weight calculation

   Attention weights are computed by taking the dot product of query and key vectors, then applying softmax.

2. Step 2: Match formula to options

   Softmax(Q x K^T) shows softmax applied to Q multiplied by the transpose of K, which is correct.

3. Final Answer:

   Softmax(Q x K^T) -> Option D

4. Quick Check:

   Attention weights = softmax(dot product) [OK]

Hint: Attention weights = softmax of query-key dot product [OK]

Common Mistakes:

• Adding Q and K instead of dot product
• Using ReLU or Sigmoid instead of softmax
• Ignoring transpose on key vector

3. Given query vector Q = [1, 0], key vectors K1 = [1, 0], K2 = [0, 1], and value vectors V1 = [10, 0], V2 = [0, 20], what is the attention output after applying softmax on Q·K^T and multiplying by values?

medium

A. [10, 0]

B. [5, 10]

C. [7.31, 5.38]

D. [0, 20]

Solution

1. Step 1: Calculate dot products Q·K1 and Q·K2

   Q·K1 = 1*1 + 0*0 = 1; Q·K2 = 1*0 + 0*1 = 0.

2. Step 2: Apply softmax to [1, 0]

   Softmax(1,0) = [e^1/(e^1+e^0), e^0/(e^1+e^0)] ≈ [0.731, 0.269].

3. Step 3: Multiply weights by values and sum

   Output = 0.731*[10,0] + 0.269*[0,20] = [7.31, 0] + [0,5.38] = [7.31, 5.38].

4. Step 4: Match to options

   The computed output [7.31, 5.38] matches [7.31, 5.38] (approximate values).

5. Final Answer:

   [7.31, 5.38] -> Option C

6. Quick Check:

   Softmax weights x values = output [OK]

Hint: Softmax weights times values gives attention output [OK]

Common Mistakes:

• Skipping softmax normalization
• Multiplying query with values directly
• Ignoring vector multiplication order

4. Identify the error in this attention weight calculation code snippet:

import numpy as np
Q = np.array([1, 2])
K = np.array([[1, 0], [0, 1]])
scores = np.dot(Q, K)
weights = np.exp(scores) / np.sum(np.exp(scores))

medium

A. Dot product should be between Q and K transpose

B. Softmax calculation is incorrect

C. Q and K should be swapped in dot product

D. No error, code is correct

Solution

1. Step 1: Check dot product dimensions

   Q is shape (2,), K is (2,2). np.dot(Q, K) results in shape (2,), but attention needs dot product with K transpose.

2. Step 2: Correct dot product usage

   Dot product should be np.dot(Q, K.T) to get scores for each key vector.

3. Final Answer:

   Dot product should be between Q and K transpose -> Option A

4. Quick Check:

   Dot product with K transpose needed [OK]

Hint: Dot product query with key transpose for scores [OK]

Common Mistakes:

• Using K instead of K transpose
• Miscomputing softmax manually
• Swapping Q and K incorrectly

5. In a transformer model, why is scaling the dot product by the square root of the key dimension important before applying softmax?

hard

A. To prevent large dot product values causing softmax to produce very small gradients

B. To increase the dot product values for better attention

C. To normalize the query vectors only

D. To reduce the number of keys processed

Solution

1. Step 1: Understand dot product scaling

   Without scaling, large dot product values can make softmax outputs very close to 0 or 1, causing gradients to vanish during training.

2. Step 2: Purpose of scaling by sqrt of key dimension

   Scaling reduces the magnitude of dot products, keeping softmax outputs more balanced and gradients healthy.

3. Final Answer:

   To prevent large dot product values causing softmax to produce very small gradients -> Option A

4. Quick Check:

   Scaling avoids gradient vanishing in softmax [OK]

Hint: Scale dot product to keep softmax gradients stable [OK]

Common Mistakes:

• Thinking scaling increases dot product values
• Believing scaling normalizes queries only
• Assuming scaling reduces keys processed