Abstract: Model stealing (MS) involves querying and observing the output of a machine learning model to steal its capabilities. The quality of queried data is crucial, yet obtaining a large amount of ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results