We have added new functionality to the Experience Training section of the platform, which will make the tool a lot easier to navigate for users. Starting with Early Access on July 28, 2021, users will have the option to turn on the Summer ‘21: New Components Experience Training (early access) and Summer ‘21: Experience Training Updates (early access) account features. To turn the features on in your account, navigate to Account Settings > Account Features, and you will find the feature under ‘Early Access Features’.
Turning on the Summer ‘21: Experience Training Updates (early access) feature will include the following updates on the Featured Snippets screen:
Deduplication: Previously, the Featured Snippets table would only show one featured snippet per search term. However, there can be multiple featured snippets for the same search term for two reasons:
- The results set can change (through changes in the Knowledge Graph or in the configuration)
- The algorithm gets retrained as it gets more feedback.
The Featured Snippets table will now dedupe on search term and featured snippet, to show one row per featured snippet per search term. Users can see all the different predictions the algorithm has made for a particular search term.
Entity Information: Users will now be able to see the entity (hyperlinked) and field that the featured snippet came from, crucial context to determine whether or not a featured snippet is correct.
Last Searched: We’ve added a column to display the most recent search time for each search term. The table is sorted by this column by default, and now you’ll be able to see it explicitly.
New vs. Completed: We’ve added a Show completed toggle, so you can see the featured snippets that have already been approved or rejected. This makes it easier for users to find completed training examples.
Rich Text: Rich text is now supported in featured snippets.
In addition we have two new components available for Experience Training from our May ‘21 Release: NLP Filters and Spell Checking. Adding NLP Filters and Spell Checking as components extends the experience training system to allow users to approve and reject these parts of the algorithm, instantly fixing issues, and providing a valuable source of training data for our algorithm. You can read more about Experience Training for NLP Filters and Spell Checking in the Community Post from our May ‘21 Release. Turn on the Summer ‘21: New Components Experience Training (early access) feature to see these new components.