CS4417/9117/9647 Midterm Review Game

In the context of indexing techniques, what is an inverted index primarily used for?

• To store documents in a sequential order

• To map terms to their locations in a document

• To compress data for efficient storage

• To encrypt data for security purposes

What is the primary goal of user query processing in information retrieval systems?

• To increase the size of the database

• To convert user queries into a format that can be understood by the system

• To delete irrelevant documents from the database

• To encrypt user data for privacy

Which of the following is an example of an information retrieval model?

• Relational Model

• Vector Space Model

• Entity-Relationship Model

• Hierarchical Model

Which component of an information retrieval system is responsible for ranking documents based on relevance?

• Indexing Engine

• Ranking Algorithm

• Query Processor

• Data Storage

What is the purpose of a stop word list in information retrieval?

• To highlight important terms in a document

• To remove common words that are unlikely to be useful for searching

• To store synonyms for query expansion

• To encrypt sensitive information

Which of the following best describes the difference between word embeddings (e.g., Word2Vec, GloVe) and sentence/text embeddings (e.g., BERT, Sentence-BERT)?

• Word embeddings capture the meaning of individual words in isolation, while sentence embeddings consider the context of words within a sentence.

• Sentence embeddings are just the sum of word embeddings, with no additional processing.

• Word embeddings are always learned using transformer-based models, while sentence embeddings are only trained using neural networks.

• Both word and sentence embeddings are interchangeable and can be used in the same way without affecting results

The backpropagation algorithm is used for

• Training the weights of a neural network

• Evaluating the performance of a neural network

• Choosing the architecture of a neural network

• None of the above

Which is true for neural networks?

• Each node computes its weighted input

• It has set of nodes and connections

• weight values can be learned

• All of the above

In an MLP used for binary classification, which activation function is commonly used in the output layer?

• Softmax

• Tanh

• Sigmoid

• ReLU

When does a shallow neural network model become a deep learning model?

• When you add more hidden layers and increase depth of neural network

• When there is higher dimensionality of data

• When the problem is an image recognition problem

• None of the options