Portfolio
Research desks
Four desks inside the think tank publish concept notes on this site. Each chapter below carries the desk’s brief, the running index of its notes, the month-by-month movement, and the documents themselves. The keyword in the path is the topic.
Computation desk
Computation desk
The computation desk writes the problem as a gate list before a simulator or a device is chosen. Width, depth, and the moment of readout are named in the note so later desks can replay the same artifact.
| Year | Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec | Sum |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2025 | — | — | 0.42 | -1.48 | 1.35 | 0.38 | -0.72 | 1.64 | 0.41 | -0.86 | 2.05 | 1.47 | 4.66 |
| 2026 | 0.86 | 0.54 | -0.33 | 1.12 | — | — | — | — | — | — | — | — | 2.19 |
The figures are the month’s movement in this desk’s note index. A dash is a month with no published note. The line above is the running sum. It is a research record, not a fund return.
Simulation desk
Simulation desk
The simulation desk keeps an exact state vector while it still fits in memory, and switches to a tensor network only when entanglement stays local. A small-circuit check has to match before a compressed run is treated as a result.
| Year | Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec | Sum |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2025 | — | — | 0.18 | 0.74 | -1.05 | 0.46 | 1.22 | -0.58 | 0.83 | 1.09 | 1.66 | 0.41 | 4.96 |
| 2026 | 1.14 | -0.27 | 0.63 | 0.48 | — | — | — | — | — | — | — | — | 1.98 |
The figures are the month’s movement in this desk’s note index. A dash is a month with no published note. The line above is the running sum. It is a research record, not a fund return.
/doc/state-vector-simulation
State-vector simulation
State-vector simulation applies each gate to the full list of amplitudes. The simulation desk uses it as the exact reference for small circuits from the quantum circuit model.
Exact simulation is the desk’s reference instrument. It is trustworthy while the state still fits in memory, and it is the check every compressed method has to match on a small circuit.
/doc/tensor-network-contraction
Tensor-network contraction
Tensor-network contraction stores a quantum state as connected tensors and sums out the bonds between them. The simulation desk uses it when a circuit from the quantum circuit model is too wide for a state vector but not fully entangled.
When entanglement stays local, the desk can replace the full amplitude list with a network of smaller tensors. The discovery in the notes is the contraction order: a poor order rebuilds the exponential state the method was meant to avoid.
Measurement desk
Measurement desk
The measurement desk treats a circuit result as samples, not as a state. Shot count, the Born probability, and the named noise model sit next to every number the desk publishes.
| Year | Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec | Sum |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2025 | — | — | 1.02 | -0.84 | -0.41 | 0.57 | 0.33 | -1.12 | 0.26 | 1.48 | 0.71 | 0.95 | 2.95 |
| 2026 | 0.22 | 0.81 | 0.16 | -0.47 | — | — | — | — | — | — | — | — | 0.72 |
The figures are the month’s movement in this desk’s note index. A dash is a month with no published note. The line above is the running sum. It is a research record, not a fund return.
/doc/measurement-and-sampling
Measurement and sampling
Measurement turns amplitudes into recorded outcomes. Sampling repeats that step. The measurement desk treats the shot count as part of the result, whether the shots come from a simulator or a device.
The desk’s finding is that a circuit does not return a state. It returns samples. Any published number is an estimate from those samples, with a count attached, or it is not a result.
/doc/noise-and-error
Noise and error
Noise is unwanted change in a quantum state while a circuit runs. The measurement desk separates that physical disturbance from ordinary sampling error, and writes both down beside every device result.
Ideal circuits and device circuits are different experiments. The desk’s finding is that an error model has to be named — depolarising, readout flips, or idle decoherence — before a gap between simulator and device is called a discovery.
Systems desk
Systems desk
The systems desk runs the classical optimiser and the quantum circuit as one loop. The shared object is the gate list. Neither side edits the other’s representation.
| Year | Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec | Sum |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2025 | — | — | -0.36 | -0.92 | 0.68 | 1.24 | 0.15 | 1.07 | -0.44 | 1.55 | 2.18 | 0.63 | 5.78 |
| 2026 | 0.77 | 1.03 | 0.29 | 0.58 | — | — | — | — | — | — | — | — | 2.67 |
The figures are the month’s movement in this desk’s note index. A dash is a month with no published note. The line above is the running sum. It is a research record, not a fund return.
Start with the circuit, then the two ways of simulating it, then readout and noise, then the hybrid loop. Every link stays on Backett Howe.
