If the machine-assisted moderation results returned by Content Moderation do not meet your expectations, you can call feedback operations to give feedback on samples.

If you use the Content Moderation API on your own platform, you can also call feedback operations to give feedback on samples whose machine-assisted moderation results do not meet your expectations. Alibaba Cloud can add the feedback data to model training and correct the moderation results in later versions. The trained model can be more adaptive to your business scenarios.

Currently, Content Moderation allows you to give feedback on image samples, video samples, and text samples. For more information, see the API reference.

Automatically add samples to a custom image library

Model training requires sufficient samples and takes some time. You can enable Content Moderation to automatically add samples to a custom image library and correct the moderation results in real time.

For the feedback on image samples, you can manage feedback-based image libraries in the Alibaba Cloud Content Moderation console. The operations to manage feedback-based image libraries are the same as those to manage user-created image libraries, except that you cannot create or delete a feedback-based image library. For more information, see Manage custom image libraries.

To automatically add samples to a custom image library, you can set the label parameter when you call feedback operations:
  • For normal images, set this parameter to normal to add these images to the whitelist.
  • For risky images, set this parameter to other values to add these images to the blacklist. We recommend that you use values that indicate risks such as porn, ad, and terrorism.