Eg: The ad inventory can be stored in the cloud. Using on device ML, a process then matches the ad to the user in order to serve the best ad to the right user.
- Are there any limitations to this? (I am planning to make an app that serves ads using TFLite and CoreML so the user privacy could be protected.)
- Can models in TFLite be updated with user data on-device, helping models stay relevant to user behavior without compromising privacy like Apple's CoreML?