Transformer architecture in NLP - Model Metrics & Evaluation

Metrics & Evaluation - Transformer architecture

Which metric matters for Transformer architecture and WHY

Transformers are often used for tasks like language translation, text classification, or question answering. The key metrics depend on the task:

• Accuracy for classification tasks, to see how many predictions are correct.
• BLEU score for translation, measuring how close the output is to human translations.
• Perplexity for language modeling, showing how well the model predicts the next word.
• Precision, Recall, and F1-score for tasks like named entity recognition or question answering, to balance correct detections and missed items.

These metrics help us understand if the Transformer is learning meaningful patterns from language data.

Confusion matrix example for a Transformer text classification task

      | Predicted Positive | Predicted Negative |
      |--------------------|--------------------|
      | True Positive (TP) = 80  | False Negative (FN) = 20 |
      | False Positive (FP) = 10 | True Negative (TN) = 90  |

      Total samples = 80 + 20 + 10 + 90 = 200

      Precision = TP / (TP + FP) = 80 / (80 + 10) = 0.89
      Recall = TP / (TP + FN) = 80 / (80 + 20) = 0.80
      F1-score = 2 * (Precision * Recall) / (Precision + Recall) = 2 * 0.89 * 0.80 / (0.89 + 0.80) ≈ 0.84
    

This matrix helps us see where the Transformer makes mistakes and how precise and complete its predictions are.

Precision vs Recall tradeoff with Transformer examples

Imagine a Transformer model detecting spam emails:

• High Precision: Few good emails are wrongly marked as spam. This means the model is careful when it says "spam." But it might miss some spam emails.
• High Recall: Most spam emails are caught. But some good emails might be wrongly marked as spam.

Depending on what matters more, we adjust the model or threshold. For spam, high precision is often preferred to avoid losing important emails.

What "good" vs "bad" metric values look like for Transformer tasks

For a Transformer doing text classification:

• Good: Accuracy above 85%, Precision and Recall above 80%, F1-score above 80%. This means the model predicts well and balances missing and wrong predictions.
• Bad: Accuracy below 60%, Precision or Recall below 50%. This means the model often guesses wrong or misses many true cases.

For language generation, a low perplexity (closer to 1) and high BLEU score (closer to 1) are signs of good performance.

Common pitfalls in Transformer model metrics

• Accuracy paradox: High accuracy can be misleading if classes are imbalanced. For example, if 95% of texts are not spam, a model always predicting "not spam" gets 95% accuracy but is useless.
• Data leakage: If test data leaks into training, metrics look unrealistically good but the model fails in real use.
• Overfitting indicators: Very high training accuracy but low test accuracy means the model memorizes training data but does not generalize.
• Ignoring task-specific metrics: Using only accuracy for translation or generation tasks misses important quality aspects.

Self-check question

Your Transformer model for fraud detection has 98% accuracy but only 12% recall on fraud cases. Is it good for production? Why or why not?

Answer: No, it is not good. Although accuracy is high, the model misses 88% of fraud cases (low recall). For fraud detection, catching fraud (high recall) is critical to avoid losses. This model would let most fraud go undetected.

Key Result

Precision, recall, and task-specific metrics like BLEU or perplexity are key to evaluate Transformer models effectively.

Practice

1. What is the main purpose of the self-attention mechanism in a Transformer model?

easy

A. To increase the number of layers in the model

B. To reduce the size of the input data

C. To convert words into numbers

D. To let the model focus on different words in the sentence at the same time

Solution

1. Step 1: Understand self-attention role

   Self-attention helps the model look at all words together and decide which words are important for each word.

2. Step 2: Match purpose with options

   To let the model focus on different words in the sentence at the same time correctly describes this as focusing on different words simultaneously, unlike other options which describe unrelated tasks.

3. Final Answer:

   To let the model focus on different words in the sentence at the same time -> Option D

4. Quick Check:

   Self-attention = focus on words together [OK]

Hint: Self-attention means focusing on all words at once [OK]

Common Mistakes:

• Thinking self-attention reduces input size
• Confusing self-attention with embedding
• Assuming it increases model layers

2. Which of the following is the correct way to describe the Transformer architecture components?

easy

A. It has encoder and decoder parts

B. It has only an encoder part

C. It uses only convolutional layers

D. It uses recurrent neural networks

Solution

1. Step 1: Recall Transformer structure

   Transformers have two main parts: encoder to process input and decoder to generate output.

2. Step 2: Compare options with structure

   It has encoder and decoder parts correctly states the presence of both encoder and decoder; others mention incorrect or unrelated components.

3. Final Answer:

   It has encoder and decoder parts -> Option A

4. Quick Check:

   Transformer = encoder + decoder [OK]

Hint: Remember: Transformer = encoder + decoder [OK]

Common Mistakes:

• Thinking Transformer has only encoder
• Confusing Transformer with CNN or RNN
• Ignoring decoder role

3. Consider this simplified Transformer encoder code snippet in Python using PyTorch:

import torch
from torch import nn

class SimpleEncoder(nn.Module):
    def __init__(self):
        super().__init__()
        self.attention = nn.MultiheadAttention(embed_dim=4, num_heads=2)
    def forward(self, x):
        attn_output, _ = self.attention(x, x, x)
        return attn_output

x = torch.rand(5, 3, 4)  # sequence length=5, batch=3, embed=4
model = SimpleEncoder()
output = model(x)
print(output.shape)

What will be the printed output shape?

medium

A. torch.Size([3, 5, 4])

B. torch.Size([5, 3, 4])

C. torch.Size([5, 4, 3])

D. torch.Size([3, 4, 5])

Solution

1. Step 1: Understand input shape and MultiheadAttention

   Input shape is (sequence length=5, batch=3, embedding=4). PyTorch MultiheadAttention expects (seq_len, batch, embed).

2. Step 2: Output shape matches input shape

   MultiheadAttention returns output with the same shape as input: (5, 3, 4).

3. Final Answer:

   torch.Size([5, 3, 4]) -> Option B

4. Quick Check:

   Output shape = input shape for MultiheadAttention [OK]

Hint: MultiheadAttention output shape matches input shape [OK]

Common Mistakes:

• Mixing batch and sequence dimensions
• Assuming output shape changes embedding size
• Confusing PyTorch input format

4. You have this Transformer decoder code snippet that throws an error:

import torch
from torch import nn

class SimpleDecoder(nn.Module):
    def __init__(self):
        super().__init__()
        self.attention = nn.MultiheadAttention(embed_dim=8, num_heads=4)
    def forward(self, tgt, memory):
        attn_output, _ = self.attention(tgt, memory, memory)
        return attn_output

tgt = torch.rand(10, 2, 8)  # target seq len=10, batch=2, embed=8
memory = torch.rand(5, 3, 8)  # memory seq len=5, batch=3, embed=8
model = SimpleDecoder()
output = model(tgt, memory)
print(output.shape)

What is the likely cause of the error?

medium

A. Sequence length mismatch between tgt and memory

B. Mismatch in embedding dimensions between tgt and memory

C. Batch size mismatch between tgt and memory

D. Number of attention heads is too high

Solution

1. Step 1: Check shapes of tgt and memory

   tgt=(10,2,8), memory=(5,3,8). Both have embedding size 8, sequence lengths differ (10 vs 5, allowed), but batch sizes differ (2 vs 3).

2. Step 2: Identify batch size mismatch

   Batch size mismatch between tgt (batch=2) and memory (batch=3) causes the RuntimeError in MultiheadAttention.

3. Step 3: Re-examine options carefully

   Embedding sizes match, sequence length mismatch is allowed, number of heads is valid. Batch size mismatch is most common error in such cases.

4. Final Answer:

   Batch size mismatch between tgt and memory -> Option C

5. Quick Check:

   Batch sizes must match for attention [OK]

Hint: Check batch sizes first when attention errors occur [OK]

Common Mistakes:

• Assuming sequence length must match
• Blaming embedding size mismatch incorrectly
• Thinking number of heads causes shape errors

5. You want to build a Transformer model for text summarization. Which combination of components is best suited for this task?

hard

A. Encoder-decoder, because summarization needs understanding input and generating output

B. Decoder only, because summarization is text generation

C. Neither encoder nor decoder, use RNN instead

D. Encoder only, because summarization needs understanding input only

Solution

1. Step 1: Understand summarization task

   Summarization requires reading input text (encoding) and producing a shorter text (decoding).

2. Step 2: Match task with Transformer parts

   Encoder-decoder architecture fits best as encoder understands input and decoder generates summary output.

3. Final Answer:

   Encoder-decoder, because summarization needs understanding input and generating output -> Option A

4. Quick Check:

   Summarization = encoder + decoder [OK]

Hint: Summarization needs both understanding and generating text [OK]

Common Mistakes:

• Choosing encoder only for generation tasks
• Choosing decoder only ignoring input understanding
• Ignoring Transformer benefits and choosing RNN