@quintumnia My feeling is that we can point to the AI guidelines ad nauseam and we're still going to be the place many new users will ask question more suited for Cross Validated, Computational Science, etc.

New users seem to tend not to read the guidelines--they just want to answers, and ask where they think seems most appropriate.

From a Stack UX perspective, I think instead of censuring them or merely migrating the question, we should make their intro to Stack wholly positive, and use the slightly out-of-scope questions as guides to the other Stacks. (i.e. AI should serve partly as a feeder to the hard science Stacks that deal with AI methods.)

This way, when new users with similar questions come to AI to ask, they may find the previous questions, read the answers, and decide to go directly to the hard science Stacks. (Alternately, those Stacks can be intimidating, so a user might want to dip their toe in the water over here, where the focus is more conceptual.)

Anyway, this is my theory

PS I also think it's quite useful for AI to take some "101" questions on such subjects because not everyone who comes to AI is necessarily aware of the types of maths utilized in the field. Getting a chance to see some of the basic problems on this Stack may strengthen the general conceptual understanding of the mathematics involved in different aspects AI.

@DukeZhou , For starters new users need to tag the question properly with the scientific term,language or technologies that they are using. As it is none of the people using those relevant scientific terms and technologies are going to find your question, as neither of existing tags are actually directing to the target audience.

@quintumnia I'm more of the opinion that it's our job to guide new users to this. Do you have any suggestions for more suitable tags?

@DukeZhou ,upto now there are tags which do not have description (briefly).Some one just propose/creates them without thinking about the target audience! thus confusing questions and new users

@quintumnia Let me ask instead, if the question were asked on Computation Science, what tags would you recommend. (having the best tags for the question might lure someone from CS to answer and point the questioner to CS for followup questions.)

@DukeZhou ,it Computational Science; and the tags i would recommend are; banded matrix,neural network,numerical analysis,optimization and linear algebra