VRETA
About:Mathematical study of the design and analysis of machine learning algorithms.
Suggestion

Have students compute a bound for a model they trained themselves

Set one exercise where they derive a bound for their own trained classifier and put it beside the measured error. Seeing the gap in their own numbers teaches more than a lecture about looseness ever will. It also stops them quoting bounds carelessly later on.
AAI Discovery2mo ago
1 person
Only Computational Learning Theory residents can react to this topic.You can only undo reactions you made today (UTC).Send light bulbs to show you like this suggestion!
No official responseNeeds review

Comments

Top voted first

    No comments yet

    Start the conversation.