Chapter 9. Personalized search
Course Content
0 / 103 completedPart 1. Modern search relevance
Chapter 1. Understanding user intent
Chapter 1. Summary
Chapter 1. How does AI-powered search work
Chapter 1. Introducing AI-powered search
Chapter 2. The structure of natural language
Chapter 2. Working with natural language
Chapter 2. Modeling domain-specific knowledge
Chapter 2. Distributional semantics and embeddings
Chapter 2. Summary
Chapter 2. Content + signals The fuel powering AI-powered search
Chapter 2. Challenges in natural language understanding for search
Chapter 3. Ranking and content-based relevance
Chapter 3. Implementing user and domain-specific relevance ranking
Chapter 3. Summary
Chapter 3. Controlling the relevance calculation
Chapter 4. Summary
Part 2. Learning domain-specific intent
Chapter 4. Crowdsourced relevance
Chapter 4. Introducing reflected intelligence
Chapter 5. Knowledge graph learning
Chapter 5. Automatically extracting knowledge graphs from content
Chapter 5. Using our search engine as a knowledge graph
Chapter 5. Using knowledge graphs for semantic search
Chapter 5. Summary
Chapter 5. Learning intent by traversing semantic knowledge graphs
Chapter 6. Using context to learn domain-specific language
Chapter 6. Query-sense disambiguation
Chapter 6. Learning related phrases from query signals
Chapter 6. Phrase detection from user signals
Chapter 6. Misspellings and alternative representations
Chapter 6. Pulling it all together
Chapter 6. Summary
Chapter 7. Interpreting query intent through semantic search
Chapter 7. Indexing and searching on a local reviews dataset
Chapter 7. An end-to-end semantic search example
Chapter 7. Summary
Part 3. Reflected intelligence
Chapter 8. Signals-boosting models
Chapter 7. Query interpretation pipelines
Chapter 8. Normalizing signals
Chapter 8. Fighting signal spam
Chapter 8. Combining multiple signal types
Chapter 8. Time decays and short-lived signals
Chapter 8. Index-time vs. query-time boosting Balancing scale vs. flexibility
Chapter 8. Summary
Chapter 9. Recommendation algorithm approaches
Chapter 9. Personalized search
Chapter 9. Summary
Chapter 9. Implementing collaborative filtering
Chapter 9. Challenges with personalizing search results
Chapter 9. Personalizing search using content-based embeddings
Chapter 10. Step 1 A judgment list, starting with the training data
Chapter 10. Learning to rank for generalizable search relevance
Chapter 10. Step 2 Feature logging and engineering
Chapter 10. Step 4 Training (and testing!) the model
Chapter 10. Step 3 Transforming LTR to a traditional machine learning problem
Chapter 10. Steps 5 and 6 Upload a model and search
Chapter 10. Rinse and repeat
Chapter 10. Summary
Chapter 11. Overcoming position bias
Chapter 11. Automating learning to rank with click models
Chapter 11. Exploring your training data in an LTR system
Chapter 11. Handling confidence bias Not upending your model due to a few lucky clicks
Chapter 11. Summary
Chapter 12. Overcoming ranking bias through active learning
Chapter 12. Exploit, explore, gather, rinse, repeat A robust automated LTR loop
Chapter 12. AB testing a new model
Chapter 12. Overcoming presentation bias Knowing when to explore vs. exploit
Chapter 12. Summary
Part 4. The search frontier
Chapter 13. Semantic search with dense vectors
Chapter 13. Search using dense vectors
Chapter 13. Getting text embeddings by using a Transformer encoder
Chapter 13. Applying Transformers to search
Chapter 13. Semantic search with LLM embeddings
Chapter 13. Cross-encoders vs. bi-encoders
Chapter 13. Natural language autocomplete
Chapter 13. Quantization and representation learning for more efficient vector search
Chapter 13. Summary
Chapter 14. Constructing a question-answering training dataset
Chapter 14. Building the reader with the new fine-tuned model
Chapter 14. Question answering with a fine-tuned large language model
Chapter 14. Fine-tuning the question-answering model
Chapter 14. Incorporating the retriever Using the question-answering model with the search engine
Chapter 14. Summary
Chapter 15. Foundation models and emerging search paradigms
Chapter 15. Multimodal search
Chapter 15. Other emerging AI-powered search paradigms
Chapter 15. Generative search
Chapter 15. Hybrid search
Chapter 15. Convergence of contextual technologies
Chapter 15. All the above, please!
Chapter 15. Summary
Appendix A. Running the code examples
Appendix A. Pulling the source code
Appendix A. Building and running the code
Appendix A. Working with Jupyter
Appendix A. Working with Docker
Appendix B. Supported search engines and vector databases
Appendix B. Swapping out the engine
Appendix B. The engine and collection abstractions
Appendix B. Adding support for additional engines