Dimension witness characterization - classic vs quantum #Generating witness scenarios

#WPoly <p> <m> <d> <o> where

• p preparations
• m measurements
• d dimension (i.e. d==2 for a bit d==3 for a trit)
• o outcomes

For d==2 two polytopes corresponding to expected values and probabilities polytope. Otherwsise only the probababilities polytope is created

Polytopes are created in folder ~/wit/scenarios/p-m-d-o ( example: ~/wit/scenarios/3-2-2-3 )

by default panda is used to find the facets of the polytope

panda pandaP.poi > panda.ieq

panda pandaE.poi > panda.ieq

filters the list of witness

select the best witnesses

The file **a<p><m><d><o>.m ** (Example: a3223.m) is created with all witnesses in matlab's cell array format #Generating witness scenarios

#Running quantum limits

• mSeeSaw computes limits using projective measurement or POVMS
• prepareNPADimNeumark - prepares momentum matrix base for a scenario
• runNPADimNeumark - run SDP for a specific witness using the base bound with prepareNPADimNeumark