Please enable JavaScript.
Coggle requires JavaScript to display documents.
AI-powered Search - Coggle Diagram
AI-powered Search
Modern Search Relevance
Introducing AI-powered search
Search capabilities
Domain-aware
Categories
Attributes
Terminology
Entities
Contextual & Personalize
User context
location
last search
profile
previous interactions
User recommendations
User classification
Domain context
Inventory
Business rules
domain-specific terminology
Query context
Other keywords
Similar keywords
Conversational
Multi-modal
Text queries
Voice queries
Images / Videos
Event monitor
Notification
Intelligent
Predict type-ahead
Spelling correction
Phrase and attribute detection
Intent classification
Conceptual searching
Assistive (moving beyond just delivery of links to delivering)
Available Actions
Answers
Search for User Intent
Discovery Engine
Recommendation
Types
Content-based recommenders
Behavior-based recommenders
Multi-modal recommenders
Search Engine
requirements
high volumn concurrency queries
delivering results in hundred of ms or less
real-time data ingestion
near-real-searching newly ingested data
cross functional of systems
massively scalable documents
strong backend
distributed systems
concurrency
data structures
operation systems
high-performance computing
specific data structure
inverted index
linear algebra
vector similarity scoring
text analysis & NLP
specific search data model
spell checking
autosuggest
faceting
text highlighting
so on
IR Continuum
Personalize search
Explicitly user input
Implicit understanding user
User-guided recommendations
Search vs Recommendation: The false dichotomy
Traditional keyword search
Personalized search
User-guide recommendation
Traditional recommendation
Take away
interpreting user intent
returning content matching that intent
Key technologies for AI-powered search
How AI-powered Search work?