Minutes from meeting
+ David re-wrote QUBO solving software
- configuration via yaml
- now uses a sparse matrix (custom implementation via triplet objects)
- if little work, benchmark against CSC sparse matrix?
- numpy solver limited to 16 qbits with dense hamiltonian (can do better than qiskit with implementation outside of it)
+ Annabel implemented track classification based on truth
- studied triplet overlap rates between fake and real tracks vs xi