Periodic Reporting for period 1 - QECANM (Quantum Estimation and Control for Advanced Noisy Metrology)
Okres sprawozdawczy: 2023-09-01 do 2025-08-31
Quantum sensors are already used to make extraordinarily precise measurements, from imaging and microscopy to tests of fundamental physics. The QECANM project set out to understand how far such sensors can be pushed under realistic conditions, where several quantities may need to be estimated at once and where environmental noise cannot be ignored. The project focused on three connected pillars: (i) understanding and exploring the limits of multi-parameter quantum estimation, including the peculiar quantum phenomenon of measurement incompatibility; (ii) deriving fundamental precision limits when probes are affected by correlated (non-independent) noise; and (iii) developing benchmarks for schemes based on time-continuous measurements that enable real-time feedback and control, especially under realistic conditions with imperfect detection.
The overarching objective was to develop theoretical and computational tools that provide reliable benchmarks for next-generation quantum sensors.
The overarching objective was to develop theoretical and computational tools that provide reliable benchmarks for next-generation quantum sensors.
Over the two-year action, QECANM delivered a coherent toolbox of results across its three pillars. These advances led to multiple peer-reviewed outputs in 2024–2025 (including PRX Quantum, Physical Review Letters, Physical Review A, Physics Letters A, and Journal of Physics A), alongside preprints under review. Further manuscripts exploiting the project’s results are in preparation.
# Multiple parameters
We developed a theoretical and numerical toolbox to quantify how measurement-stage noise degrades the information available when estimating multiple parameters, introducing the notion of susceptibility to noise for multiparameter quantum probes [1].
We clarified fundamental trade-offs in the simultaneous estimation of optical phase and loss on the same mode, highlighting the inevitable role of measurement incompatibility [2].
To support these studies, we formulated a general framework and numerical toolbox to evaluate ultimate precision bounds for bosonic Gaussian states in multiparameter settings; this was presented as a standalone contribution [3].
# Correlated noise
While single-parameter noisy quantum metrology is mature, correlated noise remains challenging. We contributed universal precision bounds that hold under correlated (non-i.i.d.) noise and are valid for arbitrary control strategies, generalising earlier independent-noise results [4]. We applied this approach to collisional thermometry, where probes interact sequentially with a common thermalising bath and become correlated, providing a discretised counterpart of time-continuous measurement models.
# Time-continuous measurements and control with noise
We produced a pedagogical resource linking discretised collisional models to realistic time-continuous measurement setups [5]. Building on this, we developed an accessible framework and an efficient tensor-network simulation pipeline to evaluate precision under inefficient detection and other Markovian noise, tailored to light–matter platforms and demonstrated on non-trivial instances [6].
We also analysed a many-body light–matter platform exhibiting a so-called boundary time-crystal behaviour; under inefficient detection we found that the quantum-enhanced scaling with particle number disappears [7].
On the control side, we studied how spectral and temporal features of single-photon pulses affect the estimation of light–matter parameters [8], and we contributed an in-depth analysis of bosonic quantum metrology with Markovian noise [9], showing that general precision bounds are directly relevant to continuous-measurement scenarios and can sometimes be saturated by such strategies.
# Research outputs (published papers and preprints)
[1] doi:10.1103/PhysRevA.110.032436
[2] doi:10.1088/1751-8121/ade516
[3] doi:10.48550/arXiv.2504.17873 (preprint)
[4] doi:10.1103/jy3v-wkcb
[5] doi:10.1016/j.physleta.2023.129260
[6] doi:10.1103/ljh3-3l4j (accepted manuscript); doi:10.48550/arXiv.2504.12399 (preprint)
[7] doi:10.48550/arXiv.2508.15448 (preprint)
[8] doi:10.1103/PhysRevA.110.043710
[9] doi:10.1103/PRXQuantum.6.020351
# Multiple parameters
We developed a theoretical and numerical toolbox to quantify how measurement-stage noise degrades the information available when estimating multiple parameters, introducing the notion of susceptibility to noise for multiparameter quantum probes [1].
We clarified fundamental trade-offs in the simultaneous estimation of optical phase and loss on the same mode, highlighting the inevitable role of measurement incompatibility [2].
To support these studies, we formulated a general framework and numerical toolbox to evaluate ultimate precision bounds for bosonic Gaussian states in multiparameter settings; this was presented as a standalone contribution [3].
# Correlated noise
While single-parameter noisy quantum metrology is mature, correlated noise remains challenging. We contributed universal precision bounds that hold under correlated (non-i.i.d.) noise and are valid for arbitrary control strategies, generalising earlier independent-noise results [4]. We applied this approach to collisional thermometry, where probes interact sequentially with a common thermalising bath and become correlated, providing a discretised counterpart of time-continuous measurement models.
# Time-continuous measurements and control with noise
We produced a pedagogical resource linking discretised collisional models to realistic time-continuous measurement setups [5]. Building on this, we developed an accessible framework and an efficient tensor-network simulation pipeline to evaluate precision under inefficient detection and other Markovian noise, tailored to light–matter platforms and demonstrated on non-trivial instances [6].
We also analysed a many-body light–matter platform exhibiting a so-called boundary time-crystal behaviour; under inefficient detection we found that the quantum-enhanced scaling with particle number disappears [7].
On the control side, we studied how spectral and temporal features of single-photon pulses affect the estimation of light–matter parameters [8], and we contributed an in-depth analysis of bosonic quantum metrology with Markovian noise [9], showing that general precision bounds are directly relevant to continuous-measurement scenarios and can sometimes be saturated by such strategies.
# Research outputs (published papers and preprints)
[1] doi:10.1103/PhysRevA.110.032436
[2] doi:10.1088/1751-8121/ade516
[3] doi:10.48550/arXiv.2504.17873 (preprint)
[4] doi:10.1103/jy3v-wkcb
[5] doi:10.1016/j.physleta.2023.129260
[6] doi:10.1103/ljh3-3l4j (accepted manuscript); doi:10.48550/arXiv.2504.12399 (preprint)
[7] doi:10.48550/arXiv.2508.15448 (preprint)
[8] doi:10.1103/PhysRevA.110.043710
[9] doi:10.1103/PRXQuantum.6.020351
QECANM advances the state of the art by replacing idealised assumptions with realistic sensing regimes and by introducing general frameworks that make previously intractable precision questions computable and comparable across platforms. The main advances and why they matter are:
- Multiparameter noise susceptibility ([1]). We introduce a principled way to quantify how measurement-stage noise degrades jointly estimable information, turning vague intuition about “sensitivity to noise” into a numerical, comparable quantity. This enables apples-to-apples assessment of competing sensing strategies when several parameters must be estimated at once—something earlier work could not do robustly.
- Ultimate bounds for Gaussian multiparameter sensing ([3]). We provide a general framework and numerical toolbox to evaluate ultimate precision bounds for bosonic Gaussian states with multiple parameters. Prior analyses were limited to special cases; our approach unifies them and makes bound evaluation systematic and scalable, opening the door to routine benchmarking in photonic and continuous-variable platforms.
- Universal limits under correlated noise ([4]). We derive precision bounds that hold even with correlated (non-i.i.d.) noise and arbitrary control strategies, generalising classic independent-noise results. This lets experimental teams tell whether a given protocol is already near the true physical limit, rather than an artefact of optimistic noise models; we illustrate this also on collisional thermometry, a discretised counterpart of continuous monitoring.
- Efficient simulation of time-continuous sensing with loss/noise ([6]). We build an accessible framework and tensor-network pipeline for time-continuous measurements (e.g. inefficient detection, Markovian noise) tailored to light–matter settings. Problems that were previously out of reach become tractable, enabling design-space exploration (readout, feedback, resource allocation) before going to the lab.
These frameworks are complemented by targeted studies that clarify real-world limits: rigorous trade-offs and incompatibility in joint phase-and-loss estimation ([2]); the loss of quantum-enhanced scaling in many-body, continuously monitored platforms under realistic inefficiencies ([7]); and analyses of how single-photon pulse shaping and bosonic noisy channels affect achievable precision and when continuous-measurement strategies can saturate fundamental bounds ([8], [9]).
# Expected impact
The primary impact is academic and theoretical: QECANM provides benchmarks, methods, and tooling that other groups can directly reuse to assess ultimate limits under realistic conditions. Because many of the results are widely applicable and the pipelines are computationally practical, they are also actionable for experiments—supporting choices of probes, measurements, and feedback in photonics, defect centres, and cold-atom platforms.
What’s needed next for uptake. (i) Wider validation with experimental datasets; (ii) packaging the toolboxes as user-friendly libraries with example notebooks; (iii) integration into community benchmark suites so different strategies can be compared on common instances.QECANM delivers benchmarks and methods that go beyond idealised models, directly addressing two bottlenecks for real-world quantum sensors: multi-parameter trade-offs and non-i.i.d. noise. The results enable designers to (a) judge whether a given sensing strategy is already near the ultimate limits under realistic noise and (b) choose control and readout strategies—especially in continuous-measurement settings—that most effectively convert quantum resources into precision. These insights are relevant to platforms such as photonics, solid-state defects and cold-atom systems.
Next steps include integrating the bounds and simulation tools into community benchmark suites and collaborating with experimental teams to validate them on specific devices.
- Multiparameter noise susceptibility ([1]). We introduce a principled way to quantify how measurement-stage noise degrades jointly estimable information, turning vague intuition about “sensitivity to noise” into a numerical, comparable quantity. This enables apples-to-apples assessment of competing sensing strategies when several parameters must be estimated at once—something earlier work could not do robustly.
- Ultimate bounds for Gaussian multiparameter sensing ([3]). We provide a general framework and numerical toolbox to evaluate ultimate precision bounds for bosonic Gaussian states with multiple parameters. Prior analyses were limited to special cases; our approach unifies them and makes bound evaluation systematic and scalable, opening the door to routine benchmarking in photonic and continuous-variable platforms.
- Universal limits under correlated noise ([4]). We derive precision bounds that hold even with correlated (non-i.i.d.) noise and arbitrary control strategies, generalising classic independent-noise results. This lets experimental teams tell whether a given protocol is already near the true physical limit, rather than an artefact of optimistic noise models; we illustrate this also on collisional thermometry, a discretised counterpart of continuous monitoring.
- Efficient simulation of time-continuous sensing with loss/noise ([6]). We build an accessible framework and tensor-network pipeline for time-continuous measurements (e.g. inefficient detection, Markovian noise) tailored to light–matter settings. Problems that were previously out of reach become tractable, enabling design-space exploration (readout, feedback, resource allocation) before going to the lab.
These frameworks are complemented by targeted studies that clarify real-world limits: rigorous trade-offs and incompatibility in joint phase-and-loss estimation ([2]); the loss of quantum-enhanced scaling in many-body, continuously monitored platforms under realistic inefficiencies ([7]); and analyses of how single-photon pulse shaping and bosonic noisy channels affect achievable precision and when continuous-measurement strategies can saturate fundamental bounds ([8], [9]).
# Expected impact
The primary impact is academic and theoretical: QECANM provides benchmarks, methods, and tooling that other groups can directly reuse to assess ultimate limits under realistic conditions. Because many of the results are widely applicable and the pipelines are computationally practical, they are also actionable for experiments—supporting choices of probes, measurements, and feedback in photonics, defect centres, and cold-atom platforms.
What’s needed next for uptake. (i) Wider validation with experimental datasets; (ii) packaging the toolboxes as user-friendly libraries with example notebooks; (iii) integration into community benchmark suites so different strategies can be compared on common instances.QECANM delivers benchmarks and methods that go beyond idealised models, directly addressing two bottlenecks for real-world quantum sensors: multi-parameter trade-offs and non-i.i.d. noise. The results enable designers to (a) judge whether a given sensing strategy is already near the ultimate limits under realistic noise and (b) choose control and readout strategies—especially in continuous-measurement settings—that most effectively convert quantum resources into precision. These insights are relevant to platforms such as photonics, solid-state defects and cold-atom systems.
Next steps include integrating the bounds and simulation tools into community benchmark suites and collaborating with experimental teams to validate them on specific devices.