AI Lesson 67 – Sentence Embeddings (SBERT) | Dataplexa

AI Course

I. AI Foundations
1. Introduction to AI 2. History of AI 3. AI vs ML vs DL 4. Search Algorithms 5. Intelligent Agents 6. Informed & Uninformed Search 7. Constraint Satisfaction 8. Adversarial Search 9. Logic in AI 10. Planning in AI 11. Knowledge Representation 12. Probabilistic AI 13. Bayesian Networks 14. Markov Models 15. AI Ethics 16. AI Applications 17. Intro to PyTorch 18. Intro to TensorFlow 19. AI Workflows 20. Data for AI
II. Machine Learning
21. Introduction to ML 22. ML Workflow 23. Supervised Learning 24. Unsupervised Learning 25. Reinforcement Learning 26. Linear Regression 27. Logistic Regression 28. SVM 29. Decision Trees 30. Random Forest 31. Gradient Boosting 32. XGBoost 33. K-Means 34. Hierarchical Clustering 35. Dimensionality Reduction 36. Feature Engineering 37. Feature Selection 38. Model Evaluation 39. Overfitting vs Underfitting 40. Cross Validation
III. Deep Learning
41. Intro to Deep Learning 42. Neural Networks 43. Activation Functions 44. Backpropagation 45. Loss & Optimizers 46. Regularization 47. CNN Basics 48. CNN Architectures 49. RNN, LSTM, GRU 50. Seq2Seq 51. Transformers 52. Attention 53. Positional Encoding 54. Training DL 55. Hyperparameter Tuning 56. Transfer Learning 57. Autoencoders 58. GAN Basics 59. GAN Applications 60. Model Serving
IV. Natural Language Processing
61. Intro to NLP 62. Text Processing 63. Tokenization 64. Stopwords & Lemmatization 65. BoW & TF-IDF 66. Word Embeddings 67. Sentence Embeddings 68. NLP Classification 69. Sequence Modeling 70. RNN in NLP 71. Transformers in NLP 72. BERT Basics 73. BERT Fine-tuning 74. RoBERTa & DistilBERT 75. Sentiment Analysis 76. NER 77. Machine Translation 78. Text Generation 79. NLP Evaluation 80. NLP Use Cases
V. Computer Vision
81. Intro to CV 82. Image Processing 83. Filters & Edges 84. Convolutions 85. Object Detection 86. Object Tracking 87. Image Segmentation 88. Face Recognition 89. Feature Detection 90. OCR 91. Image Augmentation 92. Generative Vision 93. Vision Transformers 94. Multimodal Models 95. CV Use Cases
VI. LLM Engineering
96. Intro to LLMs 97. Tokenization 98. LLM Architecture 99. Pretraining 100. Fine-tuning 101. RLHF 102. Prompt Engineering 103. Embedding Models 104. Vector Databases 105. LLM Agents 106. Guardrails & Safety 107. AI in Production 108. End-to-End Project

Sentence Embeddings (SBERT)

In the previous lesson, we learned how word embeddings represent individual words using meaning and context. However, many real-world NLP tasks involve comparing complete sentences, not just single words. Sentence embeddings solve this problem by converting entire sentences into numerical vectors.

This lesson explains what sentence embeddings are, why they are needed, and how models like Sentence-BERT (SBERT) are used in modern NLP systems.

Real-World Connection

When you search a question on Google or ask a chatbot something in your own words, the system needs to understand the meaning of the entire sentence. Even if two sentences use different words, they may mean the same thing. Sentence embeddings allow systems to measure this similarity accurately.

For example, “How can I learn AI?” and “What is the best way to study artificial intelligence?” should be treated as similar queries.

What Are Sentence Embeddings?

Sentence embeddings are dense numerical vectors that represent the meaning of an entire sentence. Unlike Bag of Words or TF-IDF, they capture semantic meaning, word order, and context.

  • Each sentence becomes a fixed-length vector
  • Similar sentences have similar vectors
  • Works well for semantic similarity tasks

Why Word Embeddings Are Not Enough

Word embeddings represent words independently. Simply averaging word vectors often loses meaning and structure. Sentence embeddings solve this by learning sentence-level representations directly.

  • Preserves sentence meaning
  • Handles different word orders
  • Understands context better

What Is Sentence-BERT (SBERT)?

Sentence-BERT is a modification of the BERT architecture designed specifically to generate meaningful sentence embeddings. It uses Siamese and triplet network structures to compare sentence pairs efficiently.

SBERT is widely used because it is fast, accurate, and produces high-quality embeddings for real-world applications.

Sentence Embedding Example Using SBERT


from sentence_transformers import SentenceTransformer

model = SentenceTransformer("all-MiniLM-L6-v2")

sentences = [
    "AI is transforming the world",
    "Artificial intelligence is changing technology",
    "I love playing football"
]

embeddings = model.encode(sentences)

print(embeddings.shape)
(3, 384)

Understanding the Output

Each sentence is converted into a 384-dimensional vector. These vectors can be compared using similarity measures such as cosine similarity to determine how close the meanings are.

Measuring Sentence Similarity


from sentence_transformers import util

similarity = util.cos_sim(embeddings[0], embeddings[1])
print(similarity)
tensor([[0.87]])

What the Similarity Score Means

A cosine similarity score close to 1 indicates high semantic similarity. The first two sentences are very similar in meaning, while the third sentence would produce a much lower similarity score.

Where Sentence Embeddings Are Used

  • Semantic search engines
  • Question answering systems
  • Chatbots and virtual assistants
  • Duplicate content detection
  • Document clustering

Practice Questions

Practice 1: What technique represents entire sentences as vectors?



Practice 2: Which model is commonly used for sentence embeddings?



Practice 3: Which metric is used to compare sentence embeddings?



Quick Quiz

Quiz 1: Sentence embeddings mainly capture what?





Quiz 2: Which method is commonly used to compare embeddings?





Quiz 3: Sentence embeddings are heavily used in which application?





Coming up next: NLP Classification Models — using embeddings to build real classifiers.

← Previous Course Index Next →