Podcast
Questions and Answers
What is the first step to request approval for a Workspace?
What is the first step to request approval for a Workspace?
What happens if you need to make changes to the Meaning Schema or induction examples?
What happens if you need to make changes to the Meaning Schema or induction examples?
What could commonly cause a submission to fail during the pre-submit checks?
What could commonly cause a submission to fail during the pre-submit checks?
What action is needed if the submission fails due to pre-submit checks?
What action is needed if the submission fails due to pre-submit checks?
Signup and view all the answers
What must be done first to create a new Meaning Schema in Intent Lab?
What must be done first to create a new Meaning Schema in Intent Lab?
Signup and view all the answers
Which tool can be used to check existing Meaning Schemas before creating a new one?
Which tool can be used to check existing Meaning Schemas before creating a new one?
Signup and view all the answers
What should you avoid enabling unless you are sure of high precision in interpretation?
What should you avoid enabling unless you are sure of high precision in interpretation?
Signup and view all the answers
What is the maximum suggested length for a description when defining a Meaning Schema?
What is the maximum suggested length for a description when defining a Meaning Schema?
Signup and view all the answers
After creating a Meaning Schema, which area should be set as Management Area?
After creating a Meaning Schema, which area should be set as Management Area?
Signup and view all the answers
What is the initial action when using the Model tool to create a new Meaning Schema?
What is the initial action when using the Model tool to create a new Meaning Schema?
Signup and view all the answers
What happens if you try to create a Meaning Schema while not in a Workspace?
What happens if you try to create a Meaning Schema while not in a Workspace?
Signup and view all the answers
What should be enabled in the left panel after creating a Meaning Schema for typical intent triggering?
What should be enabled in the left panel after creating a Meaning Schema for typical intent triggering?
Signup and view all the answers
What must be linked in the Workspace Description when answerless voting is enabled?
What must be linked in the Workspace Description when answerless voting is enabled?
Signup and view all the answers
When defining slots in Meaning Schemas, what must be specified?
When defining slots in Meaning Schemas, what must be specified?
Signup and view all the answers
Why is it important to calculate net wins versus losses in answerless voting cases relating to Trust and Safety?
Why is it important to calculate net wins versus losses in answerless voting cases relating to Trust and Safety?
Signup and view all the answers
What kind of data tag do the patterns receive when added through the Expand tool?
What kind of data tag do the patterns receive when added through the Expand tool?
Signup and view all the answers
What is the primary purpose of using slots in Meaning Schemas?
What is the primary purpose of using slots in Meaning Schemas?
Signup and view all the answers
Which of the following best describes the 'Expand tool'?
Which of the following best describes the 'Expand tool'?
Signup and view all the answers
What should be done after defining the slots for your Meaning Schema?
What should be done after defining the slots for your Meaning Schema?
Signup and view all the answers
In the context of answerless voting, what is the key reason for not allowing certain answers?
In the context of answerless voting, what is the key reason for not allowing certain answers?
Signup and view all the answers
What happens when a pattern is marked as good?
What happens when a pattern is marked as good?
Signup and view all the answers
What should be done with patterns after they are added to the Meaning Schema?
What should be done with patterns after they are added to the Meaning Schema?
Signup and view all the answers
What feature suggests patterns based on previously marked good patterns?
What feature suggests patterns based on previously marked good patterns?
Signup and view all the answers
Which of the following describes the process of adding multiple patterns at once?
Which of the following describes the process of adding multiple patterns at once?
Signup and view all the answers
What is the outcome of marking a pattern as bad?
What is the outcome of marking a pattern as bad?
Signup and view all the answers
What is the significance of KScorer salient terms?
What is the significance of KScorer salient terms?
Signup and view all the answers
How does the system identify patterns that should trigger a search stack?
How does the system identify patterns that should trigger a search stack?
Signup and view all the answers
What must be done if a new Meaning Schema is created?
What must be done if a new Meaning Schema is created?
Signup and view all the answers
What is the main purpose of the KE SxS eval?
What is the main purpose of the KE SxS eval?
Signup and view all the answers
What should be included in the KE SxS eval to enhance its effectiveness?
What should be included in the KE SxS eval to enhance its effectiveness?
Signup and view all the answers
Which eval is specifically designed to ensure that implementing the Meaning Schema does not break anything?
Which eval is specifically designed to ensure that implementing the Meaning Schema does not break anything?
Signup and view all the answers
How can the load test be initiated?
How can the load test be initiated?
Signup and view all the answers
What is the function of the 'Modify' option in the Workspace description?
What is the function of the 'Modify' option in the Workspace description?
Signup and view all the answers
What does the tags section in the Workspace description accomplish?
What does the tags section in the Workspace description accomplish?
Signup and view all the answers
What happens to the tags once an eval is completed?
What happens to the tags once an eval is completed?
Signup and view all the answers
Which statement describes the purpose of the Workspace Review tool?
Which statement describes the purpose of the Workspace Review tool?
Signup and view all the answers
Study Notes
Modeling Meaning Schemas
- Modeling a Meaning Schema is easy using Intent Lab
- General SLS team approach is outlined
- Detailed modeling for complex schemas may need specialized tools/evaluation
- Refer to
go/meaning/concepts
for in-depth information
- Refer to
Creating a Meaning Schema
- Check existing schemas in the Meaning Schema catalog
- Use Query Examples Tool or Query Debugger to see if intended queries match existing schemas
- Use Model Tool or Meaning Explorer Tool to review slots of a schema
- If a schema does not exist, use Intent Lab (IL)
- Familiarize yourself with IL tools and associated concepts
- Refer to the Glossary for clarification
Workspace Creation
- Work within an existing workspace to create a new schema
- Refer to the link for workspace creation workflow
Model Tool
- Create new Meaning Schema (MS) using the Model Tool, and provide details
- Give relevant name, description, and define slots
- Understand Google's meaning definition (refer to
go/meaning
)
- Steps to create new MS:
- Navigate to the Model tool
- Input new MS name
- Use "Create NameOfSchema" button
- Provide a description (typically a sentence)
- Select "Search" as the Management Area
- Wait for Management area approval
- Click "Create"
- Verify completion of previous steps
Loose Parsing and Answerless Voting
- Enable Loose Parsing in the left panel for Search
- Enable offline parsing when enabling Loose Parsing for a MS
- Disable answerless voting unless needed (e.g., high precision is required) and link Workspace Description with a justification
- Add slots for location etc., using KG collections or MIDs
- Specify required/optional slots
Adding Patterns (Expand Tool)
- Use the Expand tool to manage patterns matching a meaning schema
- Define the slots of the meaning Schema before using the Expand Tool to add patterns
- Add multiple patterns on separate lines
- Mark patterns as "good" (triggering the schema in a search stack) or "bad" (blocking the pattern)
- Use "Fetch and Rank" functionality to suggest patterns
- Predicted patterns are light green and automatically trigger the schema
KScorer Salient Terms
- Generate KScorer salient terms, for new schemas, to be submitted to the system
- Salient terms are important for the KScorer Model (Google's ranking algorithm)
Evaluations
- Run evaluations to ensure Search functionality isn't compromised
- Use KE Evals and SxS evals; Refer to
Overview: Evals and KE Quality
for more details on these
- Use KE Evals and SxS evals; Refer to
-
KE SxS
orKE SxS with fallback fulfillment
evaluates the schema against a production version of Search
Workspace Review
- Navigate to the Workspace Review tool to examine changes
- Patterns from the expand tool are shown using the query examples
- Update workspace description by pressing "Modify"
- Add links to evaluations for clarity
Submission
- After approval, submit the schema
- Pre-submission validation checks exist
- Issues (like invisible diffs) might arise, address them, if faced
- Use resources (e.g.,
go/pp-where-is-my-cl
,go/pp-where-is-my-cl-dashboard
) to see when the changes will be available in production
Studying That Suits You
Use AI to generate personalized quizzes and flashcards to suit your learning preferences.
Related Documents
Description
This quiz covers the process of modeling Meaning Schemas using Intent Lab and explores tools and approaches for schema creation. Participants will learn about the importance of existing schemas, how to utilize tools like the Query Examples Tool and Model Tool, and steps to create new schemas. Familiarity with these resources will enhance your understanding of complex meaning representations.