i can see arguments that perhaps it is { more art than science / too application-dependent / ... }, but i suspect there are some good partial answers out there. an indicative first guess at what i have in mind is ≈ the atomic norms paper (https://t.co/Hpm
16,608 followers
3,091 followers
RT @sam_power_825: @iugoaoj https://t.co/HpmGol5zGA is a nice paper which gives some good examples of sparse-type regularisers. these 2 boo…
16,608 followers
@iugoaoj https://t.co/HpmGol5zGA is a nice paper which gives some good examples of sparse-type regularisers. these 2 books { https://t.co/Ipy6SFm7wI, https://t.co/deVkbkFDAy } are about optimisation methods which were (to some extent) designed to work nice
1,206 followers
29 followers
a nice generalised view of the linear inverse problem: http://t.co/j1IdOfai