Moreover, Evolution Gym is easy to use.

On stationary problems with low-d structure, the magnitude of5 improvement is large (Figure 1). Now, as we show, real-world problems can be more complicated and while the6 improvement overREMBO remained large,local-search methods werehighly competitive.

Submodularity is a fundamental concept in combinatorial optimization, usually associated with discrete setfunctions [Fujishige, 1991]. Assubmodular functions formalize theintuitivenotion of diminishing returns, and thus provide a useful structure, they appear in a wide range of modernmachine learning applications including various forms of data summarization [Lin and Bilmes,2012, Mirzasoleiman et al., 2013], influence maximization [Kempe et al., 2003], sparse and deeprepresentations [Balkanski etal.,2016,

To mitigate the threat, existing solutions attempted to search for backdoor triggers, which can be time-consuming when handling a large search space.