Tensor algorithm auto-scheduling in MLIR

MLIR group,

Is there a project or research that does tensor based algorithm auto-scheduling (like Halide, TVM, Tiramisu) within MLIR (e.g., a dialect)?

For schedule I mean the computation order of a nested for loop, like in the DSL’s mentioned above in which they separate schedule from algorithm computation. I know MLIR has polyhedral optimization but this is done at a lower level in which the for loops are specified. I am referring to auto-scheduling on vectorized tensor computations (no for loops).