Word similarity and analogies in NLP - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to load a pre-trained word embedding model using gensim.

NLP

from gensim.models import KeyedVectors
model = KeyedVectors.load_word2vec_format('[1]', binary=True)

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

Aword2vec.txt

Bfasttext.vec

Cglove.6B.100d.txt

DGoogleNews-vectors-negative300.bin

Attempts:

3 left

💡 Hint

Common Mistakes

Using a text file instead of a binary file for Word2Vec.

Confusing GloVe or FastText files with Word2Vec binary format.

The GoogleNews-vectors-negative300.bin file is a common pre-trained Word2Vec binary model used with gensim.

2fill in blank

medium

Complete the code to find the top 3 words most similar to 'king' using the model.

NLP

similar_words = model.most_similar('[1]', topn=3)

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

Aking

Bqueen

Cman

Droyal

Attempts:

3 left

💡 Hint

Common Mistakes

Passing a similar word like 'queen' instead of the target word 'king'.

Passing a concept or unrelated word.

The method most_similar expects the target word as input, here 'king'.

3fill in blank

hard

Fix the error in the analogy code to find the word that fits: 'man' is to 'king' as 'woman' is to ____.

NLP

result = model.most_similar(positive=['king', '[1]'], negative=['man'], topn=1)

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

Awoman

Blady

Cqueen

Dprince

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'queen' in positive instead of 'woman'.

Confusing positive and negative lists.

In the analogy, 'woman' is the positive word to add, while 'man' is subtracted.

4fill in blank

hard

Fill both blanks to create a dictionary of words and their similarity scores to 'computer', filtering only words with similarity greater than 0.7.

NLP

similarity_dict = {word: [1] for word, score in model.most_similar('[2]', topn=10) if score > 0.7}

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

Ascore

Bword

C'computer'

D'laptop'

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'word' instead of 'score' as dictionary value.

Using 'laptop' instead of 'computer' as target word.

We want the similarity score as value and 'computer' as the target word.

5fill in blank

hard

Fill all three blanks to create a list of words from the model's vocabulary that have length greater than 5 and contain the letter 'a'.

NLP

filtered_words = [[1] for [2] in model.index_to_key if [3]]

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

Aword

C'a' in word and len(word) > 5

Dlen(word) > 5

Attempts:

3 left

💡 Hint

Common Mistakes

Using different variable names inconsistently.

Using only length condition without checking for 'a'.

The list comprehension iterates over words, selects words with 'a' and length > 5, and collects the word itself.