Welcome to squint, a differentiable framework for the study and design of quantum metrology & sensing protocols!

What can it do?

• Differentiable quantum dynamics for qubit, qudit, and Fock/photon-number systems
• Built on the JAX ecosystem for automatic differentiation and GPU hardware acceleration
• Compute Fisher information and other fundamental metrics in quantum metrology with ease
• Benchmark realistic protocols with noise and loss channels relevant to diverse device architectures

Installation

Simply clone the repo locally,

git clone https://github.com/benjimaclellan/squint
cd squint

Use uv for the package management. Installation instructions are here.

Create a virtual environment with all the correct dependencies and activate it,

uv sync
source .venv/bin/activate

Example

from squint.circuit import Circuit
from squint.simulator.tn import Simulator
from squint.ops.base import Wire
from squint.ops.dv import DiscreteVariableState, HGate, RZGate
from squint.utils import print_nonzero_entries, partition_op

# let's implement a simple one-qubit circuit for phase estimation;
#          _____     _________     _____      _____
# |0 > --- | H | --- | Rz(Φ) | --- | H | ---- | / |====
#          -----     ---------     -----      -----

wire = Wire(dim=2, idx=0)  # qubit with dim=2
circuit = Circuit()
circuit.add(DiscreteVariableState(wires=(wire,), n=(0,)))
circuit.add(HGate(wires=(wire,)))
circuit.add(RZGate(wires=(wire,), phi=0.0 * jnp.pi), "phase")
circuit.add(HGate(wires=(wire,)))

params, static = partition_op(circuit, "phase")
sim = Simulator.compile(static, params, optimize="greedy").jit()

# Calculate metrics important to quantum metrology & sensing protocols
# the quantum state and its gradient
psi = sim.amplitudes.forward(params)      # |ψ(φ)⟩
dpsi = sim.amplitudes.grad(params)        # ∂|ψ(φ)⟩/∂φ

# Probabilities and their gradients
p = sim.probabilities.forward(params)     # p(s|φ)
dp = sim.probabilities.grad(params)       # ∂p(s|φ)/∂φ

qfi = sim.amplitudes.qfim(params)       # Quantum Fisher Information
cfi = sim.probabilities.cfim(params)    # Classical Fisher Information

Roadmap

Note

This package is still under heavy development, so expect major breaking changes in future versions.

squint aims to provide high-performance, flexible simulation tooling for quantum metrology, sensing, and photonics systems. Future plans including adding the ability to specify operations acting on internal and coupled degrees-of-freedom of optical information carriers, abstract POVM and post-selected measurements, and addition figures-of-merit relevant to quantum optical and metrology protocols.

Using & Contributing

If this package is useful for your work, and you have feature suggestions and feedback, please reach out by email or open a discussion on Github!

Acknowledgments

The authors of squint acknowledge the kind support of Ki3 Photonics Technologies, the Perimeter Institute Quantum Intelligence Lab, the Institute for Quantum Computing, and the NSERC Vanier Program.

Citing

If you found this package useful, please consider citing our work;

@article{maclellan2024endtoend,
  title={End-to-end variational quantum sensing}, 
  author={Benjamin MacLellan and Piotr Roztocki and Stefanie Czischek and Roger G. Melko},
  year={2024},
  eprint={2403.02394},
  archivePrefix={arXiv},
  primaryClass={quant-ph}
}