Hybrid search strategies in Prompt Engineering / GenAI - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

Understanding Hybrid Search Strategy Components

Which two main components are combined in a hybrid search strategy to improve search results?

A Decision trees and random forests

B Vector similarity search and keyword-based search

C Clustering and dimensionality reduction

D Reinforcement learning and supervised learning

💡 Hint

Think about combining semantic understanding with exact matching.

Hybrid search strategies combine vector similarity search, which captures semantic meaning, with keyword-based search, which matches exact words, to improve search accuracy and relevance.

❓ Model Choice

intermediate

Choosing a Model for Vector Search in Hybrid Systems

Which model type is best suited to generate embeddings for vector similarity search in a hybrid search system?

A Transformer-based language model trained on text

B Convolutional Neural Network (CNN) trained on images

C Linear regression model

D K-means clustering model

💡 Hint

Consider which model understands text semantics well.

Transformer-based language models like BERT or GPT generate meaningful text embeddings that capture semantic relationships, making them ideal for vector similarity search.

❓ Metrics

advanced

Evaluating Hybrid Search Performance

Which metric best measures the balance between precision and recall in hybrid search results?

A Mean Squared Error (MSE)

B Log Loss

C F1 Score

D Accuracy

💡 Hint

This metric combines precision and recall into one number.

F1 Score is the harmonic mean of precision and recall, making it suitable to evaluate how well hybrid search balances finding relevant results and avoiding irrelevant ones.

🔧 Debug

advanced

Debugging Hybrid Search Code Output

What is the output of this Python code snippet simulating a hybrid search result ranking? ```python vector_scores = [0.9, 0.75, 0.6] keyword_scores = [0.8, 0.85, 0.5] combined_scores = [v * 0.6 + k * 0.4 for v, k in zip(vector_scores, keyword_scores)] sorted_indices = sorted(range(len(combined_scores)), key=lambda i: combined_scores[i], reverse=True) print(sorted_indices) ```

A [0, 1, 2]

B [1, 0, 2]

C [2, 1, 0]

D [0, 2, 1]

💡 Hint

Calculate combined scores and sort descending.

Combined scores are [0.9*0.6+0.8*0.4=0.54+0.32=0.86, 0.75*0.6+0.85*0.4=0.45+0.34=0.79, 0.6*0.6+0.5*0.4=0.36+0.2=0.56]. Sorted descending indices are [0,1,2].

❓ Hyperparameter

expert

Tuning Weight Parameters in Hybrid Search

In a hybrid search system combining vector and keyword scores as `final_score = alpha * vector_score + (1 - alpha) * keyword_score`, which alpha value is most likely to emphasize semantic similarity over exact keyword matches?

A 0.0

B 0.5

C 0.1

D 0.9

💡 Hint

Higher alpha means more weight on vector scores.

Alpha close to 1 (like 0.9) gives more weight to vector similarity, emphasizing semantic meaning over keyword matching.