Why embeddings capture semantic meaning in NLP - Test Your Understanding

Practice - 5 Tasks

1fill in blank

easy

Complete the code to create a simple word embedding using a dictionary.

NLP

word_embeddings = {'cat': [0.1, 0.3, 0.5], 'dog': [1]

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

A[0.2, 0.4, 0.6]

B[1, 2, 3]

C[0.5, 0.5, 0.5]

D[0, 0, 0]

Attempts:

3 left

💡 Hint

Common Mistakes

Choosing a vector with different length or scale.

The embedding for 'dog' should be a vector similar in size and scale to 'cat' to capture semantic meaning.

2fill in blank

medium

Complete the code to calculate cosine similarity between two embeddings.

NLP

import numpy as np

def cosine_similarity(vec1, vec2):
    dot_product = np.dot(vec1, vec2)
    norm1 = np.linalg.norm(vec1)
    norm2 = np.linalg.norm(vec2)
    return dot_product / [1]

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

Anorm1 + norm2

Bnorm1 - norm2

Cnorm1 * norm2

Dnorm1 / norm2

Attempts:

3 left

💡 Hint

Common Mistakes

Using addition or subtraction instead of multiplication.

Cosine similarity divides the dot product by the product of the norms of the two vectors.

3fill in blank

hard

Fix the error in the code that trains a simple embedding layer in PyTorch.

NLP

import torch
import torch.nn as nn

embedding = nn.Embedding(num_embeddings=10, embedding_dim=3)
input_indices = torch.tensor([1, 2, 3])
output = embedding([1])
print(output)

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

Atensor

Binput

Cindices

Dinput_indices

Attempts:

3 left

💡 Hint

Common Mistakes

Using undefined variable names.

The embedding layer expects a tensor of indices as input, here named 'input_indices'.

4fill in blank

hard

Fill both blanks to create a dictionary comprehension that maps words to their embedding lengths if length is greater than 3.

NLP

words = ['apple', 'cat', 'banana', 'dog']
lengths = {word: [1] for word in words if len(word) [2] 3}

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

Alen(word)

B>

C<

Dword

Attempts:

3 left

💡 Hint

Common Mistakes

Using the word itself instead of its length.

Using '<' instead of '>'.

The comprehension calculates the length of each word and filters words longer than 3 characters.

5fill in blank

hard

Fill all three blanks to create a dictionary of words and their embeddings filtered by similarity score greater than 0.5.

NLP

similarities = {'apple': 0.7, 'cat': 0.4, 'banana': 0.8, 'dog': 0.3}
embeddings = {'apple': [0.1, 0.2], 'cat': [0.3, 0.4], 'banana': [0.5, 0.6], 'dog': [0.7, 0.8]}
filtered = [1]: [2] for [3], score in similarities.items() if score > 0.5}

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

Aword

Bembeddings[word]

Dscore

Attempts:

3 left

💡 Hint

Common Mistakes

Using score as key or value incorrectly.

Not matching variable names.

The dictionary comprehension uses the word as key, its embedding as value, and filters by score > 0.5.