Though the Machine Learning needs a data set to train upon! If we want to train a ML algo for a specific software i.e Paint then its only one application let say we have many paint application from different vendors and API that make data set of 100 paint application even more, then what would be the target function in this case? might be this question make no sense at all? let say if we want to learn the functionality of different Paint applications i.e their drawing, rendering fx or buttons, layers and behaviors as each program has different language and framework behind, in the end this would be very complex to having such kind of ML Algo. As AlfaGo Zero (Deepmind ALGO from Google) suggest to beat supper human in chess, chinese chess and in Go, then why not we can learn different rules for the given programs? and if given ML algo able to do so then might we have standard API for all the vendors?
- Badges
- Science topic