Text experiements: RISE method parameters range

[elboyran]

Define for which parameters would RISE be evaluated and how the continuously changing parameter range would be defined.

See the dianna's implementation of RISE and examples of using RISE on text and on images: binaryMNIST and imagnet.

The outcome of this issue should be:
<parameter>: <range>,
where
<range> = <start_value>:<step>:<end_value>
for every methods' parameter (if more than 1).
See also the dashboard for modifiable parameters.

Stems from dianna-ai/dianna#445 and #187 (see for Practicalities).

[cpranav93]

Could it be interesting to split the parameters of importance and their ranges for images and text? An interesting study could be to see if the parameters and their range vary with model/case-study.

[elboyran]

@cpranav93 good suggestion for a study. I propose to potentially make a follow up issue if the time permits to answer such a question.
As for grouping parameters based on importance, not sure they are enough number to make grouping make sense. Whoever picks up this issues should verify if it makes sense.