Are you only accepting machine learning solutions for this challenge?
I noticed that you included a clause in the Unbreak My Seismic challenge that states that only ML / DL solutions are allowed, whereas there is no such clause in this challenge. Is that intentional?
For the Parallel Perspectives Challenge we are only accepting machine learning or deep learning submissions. For this Challenge your job will be to build a machine learning or deep learning model that can unscramble these traces and restore the original line.
has the solution need to be pure machine learning?
i can see that the task is related permutation/sorting and a key component is similarity matching (between 2 strips or blocks for related 2d jigsaw puzzle)
Say i use deep learning/ML to learn similarity matching and it is so accurate that the top-1 neighbours neighbour is 100%, then i can simply unscramble by linking to the top neighbours up(e.g. greedy search + backtracking).
Is this considered a valid solution as it uses DL/ML + heuristics/combinatorics?
It sounds like you have a solution in mind that combines deep learning or machine learning and leverages a bit of a heuristic or combinatoric approach. We will consider the submission if it is primarily deep learning or machine learning based and scores well on the hold out data.