arxiv: 2604.25193 · v1 · submitted 2026-04-28 · ⚛️ physics.optics · math.OC· physics.comp-ph· physics.data-an· quant-ph

Recognition: unknown

Adaptive Sensing beyond Non-Adaptive Information Limits: End-to-End Co-Design of Geometry, Policy, and Inference

Arvin Keshvari , William Tuxbury , Zin Lin

Authors on Pith no claims yet

Pith reviewed 2026-05-07 15:47 UTC · model grok-4.3

classification ⚛️ physics.optics math.OCphysics.comp-phphysics.data-anquant-ph

keywords adaptive sensingjoint optimizationdynamic programminginverse designsensor co-designphotonic metasensorPOMDPinformation theory

0 comments

The pith

Sensors achieve large performance gains by jointly optimizing their physical geometry and adaptive measurement policy instead of designing them separately.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper argues that because any information lost at the hardware stage cannot be recovered later, the physical design of a sensor must be optimized together with the rule that decides what to measure next. It introduces joint dynamic programming to perform this co-optimization over continuous geometry parameters and a Bellman-optimal policy. The method uses exact gradients from the envelope theorem and a hierarchy of relaxations to handle everything from small problems to large photonic arrays. In three demonstrations the joint approach improves results by factors of 2.8, 11.3 and 123 compared with standard baselines that optimize geometry without the policy or vice versa. A reader should care because it changes the design workflow for any sensor that will be used adaptively over its lifetime.

Core claim

We formulate this co-design as joint dynamic programming (joint-DP): a single optimization over the continuous hardware geometry and a Bellman-optimal adaptive measurement policy. The outer hardware gradient is computed by differentiable dynamic programming with a sharp Bellman maximum, which the envelope theorem makes exact and bias-free, and a relaxation hierarchy carries the common framework from small discrete POMDPs to 10^5-pixel photonic topologies. Across three case studies, joint-DP beats the natural baseline of its community by a large factor: on a radar beam-search POMDP, classical information-bound-guided geometry selection loses 2.8 times in attainable adaptive value; on a qubit,

What carries the argument

joint dynamic programming (joint-DP), a single optimization over continuous hardware geometry and Bellman-optimal adaptive measurement policy that uses differentiable dynamic programming with a sharp Bellman maximum for exact outer gradients via the envelope theorem.

If this is right

Hardware designed without regard to the adaptive policy loses up to 2.8 times the attainable value in beam-search tasks.
Co-optimized geometries for qubit sensors cut mean-squared error by more than an order of magnitude over Bayesian Cramér-Rao methods.
Photonic metasensors reach 123-fold error reduction when geometry and policy are designed together versus random selection.
The same joint framework applies across discrete POMDPs and continuous large-scale photonic systems.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

Sensor manufacturers may need to simulate full adaptive policies during the hardware design phase rather than after fabrication.
The approach could reveal counter-intuitive geometries that only become optimal when the policy is allowed to adapt to the hardware's specific strengths.
Extending the method to multi-objective trade-offs such as power consumption and robustness would be a natural next step.
Similar co-design principles might apply to other systems where hardware is fixed but decisions continue over time, such as in robotics or communications.

Load-bearing premise

The outer hardware gradient computed by differentiable dynamic programming with a sharp Bellman maximum is exact and bias-free via the envelope theorem, and the relaxation hierarchy scales to 10^5-pixel photonic topologies without significant approximation error.

What would settle it

Compute the true joint optimum by exhaustive search on a small discrete POMDP and verify whether the joint-DP solution reaches the same value within floating-point tolerance.

Figures

Figures reproduced from arXiv: 2604.25193 by Arvin Keshvari, William Tuxbury, Zin Lin.

**Figure 1.** Figure 1: Scene illustrations for the two case studies. view at source ↗

**Figure 2.** Figure 2: Case Study C overview. A 2 × 2 lattice of identical cross-waveguide scatterers (cells 1–4), each carrying the same inverse-designed binary-permittivity pattern in its central design region. Two coherent broadband light pulses enter at the left edge: an action pulse carrying the adaptive input phases sk = (ϕ (k) 1 , ϕ(k) 2 ) at epoch k, and a fixed-phase reference pulse. External stimulus fields (e.g., elec… view at source ↗

**Figure 3.** Figure 3: Final optimized binary permittivity of the 300 view at source ↗

**Figure 1.** Figure 1: Reachable count-tuple DAG for a Ramsey-like DP at view at source ↗

read the original abstract

Inverse design has made vast physical parameter spaces a substrate for emergent behavior. In sensing, the stakes of this principle are sharpest at the analog-to-digital boundary, where any information the hardware fails to capture is information no downstream algorithm can recover; hardware optimization alone is therefore not enough, and the geometry must be co-designed with a rule for what to measure next. We formulate this co-design as \emph{joint dynamic programming} (joint-DP): a single optimization over the continuous hardware geometry and a Bellman-optimal adaptive measurement policy. The outer hardware gradient is computed by differentiable dynamic programming with a sharp Bellman maximum, which the envelope theorem makes exact and bias-free, and a relaxation hierarchy carries the common framework from small discrete POMDPs to $10^5$-pixel photonic topologies. Across three case studies, joint-DP beats the natural baseline of its community by a large factor: on a radar beam-search POMDP, classical information-bound-guided geometry selection loses $2.8\times$ in attainable adaptive value; on a superconducting-qubit flux sensor, joint-DP reduces deployed mean-squared error by $11.3\times$ over the joint Bayesian Cram\'er--Rao baseline, with both geometries numerically co-optimized under their respective objectives; and on a $90{,}000$-pixel photonic metasensor, joint-DP via the Bayesian Fisher-information-matrix surrogate reduces deployed mean-squared error by $123\times$ relative to a randomized baseline. For any sensor whose hardware is designed once but whose policy runs for the device's lifetime, joint optimization of hardware and policy is the minimum principled procedure.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

Joint DP co-design of sensor geometry and policy reports large gains but rests on whether envelope-theorem gradients stay unbiased through the relaxations and sharp Bellman max.

read the letter

The main takeaway is that this paper frames sensor co-design as a single dynamic program over continuous geometry and an adaptive policy, then uses differentiable DP plus the envelope theorem to optimize the hardware parameters without an inner loop. They scale it with a relaxation hierarchy to handle everything from small POMDPs up to 90,000-pixel metasensors, and the three examples show clear separation from the baselines: 2.8x on radar beam search, 11.3x MSE reduction on the qubit flux sensor, and 123x on the photonic case relative to randomized design. The argument that joint optimization is the minimum principled move when hardware is fixed once and the policy runs for the lifetime of the device is direct and hard to dismiss on its own terms. The formulation pulls together inverse design ideas with POMDP machinery in a way that feels like a genuine synthesis rather than a routine extension. The case studies are useful because they span radar, superconducting qubits, and large photonic arrays, which gives some sense of where the method might apply. The soft spot is exactly the one the stress-test flags: once you replace the value function with a relaxed surrogate and keep a sharp max in the Bellman step, it is not automatic that the envelope theorem still supplies an unbiased outer gradient for the geometry variables. Any residual non-differentiability or looseness in the hierarchy would feed straight into the hardware update and could undercut the optimality claim. The numerical multipliers are large enough that I would want to see explicit checks on gradient accuracy and how tight the relaxations actually are before treating the results as settled. This is for people already working on end-to-end sensing design in optics, quantum information, or radar who want a principled way to stop treating geometry and policy as separate stages. A reader who cares about co-optimization or differentiable DP would get concrete value from the setup even if they end up modifying the relaxations. It deserves a serious referee because the core problem is real and the proposed machinery is non-obvious. I would send it to review with the expectation that the main questions will be on the gradient exactness and the scaling behavior of the hierarchy.

Referee Report

2 major / 2 minor

Summary. The paper claims that joint dynamic programming (joint-DP) enables end-to-end co-design of sensor hardware geometry and adaptive Bellman-optimal policies, with exact bias-free outer gradients obtained via differentiable dynamic programming, a sharp Bellman maximum, and the envelope theorem, plus a relaxation hierarchy that scales the approach from small POMDPs to 10^5-pixel photonic devices. Three case studies report large gains over baselines: 2.8× adaptive value on a radar beam-search POMDP, 11.3× lower MSE on a superconducting-qubit flux sensor, and 123× lower MSE on a 90,000-pixel photonic metasensor using a Bayesian Fisher-information surrogate.

Significance. If the methodological claims hold, the work would be significant for inverse design and adaptive sensing: it supplies a principled procedure for co-optimizing fixed hardware with its lifetime policy, moving beyond separate non-adaptive information bounds. The reported numerical improvements are large and the scaling to large photonic topologies is a concrete strength; the framework could influence design practice in optics and quantum sensing where hardware is fabricated once.

major comments (2)

[Abstract and the section deriving the hardware gradient via the envelope theorem] The central optimality argument for joint-DP rests on the outer hardware gradient being exact and unbiased. The abstract states this follows from differentiable DP with a sharp Bellman maximum plus the envelope theorem, yet the relaxation hierarchy used to reach 10^5-pixel topologies replaces the value function with a surrogate; it is not shown whether the envelope theorem continues to deliver an unbiased gradient once the sharp maximum is relaxed. This point is load-bearing for the claim that joint optimization is the minimum principled procedure.
[relaxation hierarchy and photonic metasensor case study] § on the relaxation hierarchy: the manuscript must demonstrate that approximation error in the hierarchy does not propagate into geometry updates in a way that invalidates the reported gains (e.g., via a controlled comparison of relaxed vs. exact gradients on a small instance before scaling). Without this, the 123× improvement on the photonic metasensor cannot be confidently attributed to the joint-DP procedure rather than surrogate artifacts.

minor comments (2)

[Notation and preliminaries] Notation for the joint state-action space and the precise definition of the 'sharp Bellman maximum' should be introduced earlier and used consistently across the three case studies.
[Case studies] The radar and qubit baselines are described as 'natural' or 'joint Bayesian Cramér–Rao'; a short table comparing the exact optimization objectives used for each baseline versus joint-DP would improve reproducibility.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading and constructive comments. The points raised about the unbiasedness of the hardware gradient under relaxation and the need for validation of the hierarchy are important for rigor. We respond to each major comment below and have revised the manuscript to address them.

read point-by-point responses

Referee: [Abstract and the section deriving the hardware gradient via the envelope theorem] The central optimality argument for joint-DP rests on the outer hardware gradient being exact and unbiased. The abstract states this follows from differentiable DP with a sharp Bellman maximum plus the envelope theorem, yet the relaxation hierarchy used to reach 10^5-pixel topologies replaces the value function with a surrogate; it is not shown whether the envelope theorem continues to deliver an unbiased gradient once the sharp maximum is relaxed. This point is load-bearing for the claim that joint optimization is the minimum principled procedure.

Authors: We thank the referee for identifying this key subtlety in the argument. The relaxation hierarchy approximates only the inner maximization over policies while preserving differentiability of the outer hardware-to-value mapping; the envelope theorem is applied to the differentiable DP operator prior to relaxation, and the sequence of relaxations (temperature-annealed soft-max) is constructed so that the limit recovers the sharp case with unbiased gradients. To make the argument fully explicit, we have added a new theorem with proof sketch in the revised methods section showing that the hardware gradient remains unbiased under the hierarchy. We also include a brief numerical check on a toy instance confirming gradient agreement to machine precision in the limit. revision: yes
Referee: [relaxation hierarchy and photonic metasensor case study] § on the relaxation hierarchy: the manuscript must demonstrate that approximation error in the hierarchy does not propagate into geometry updates in a way that invalidates the reported gains (e.g., via a controlled comparison of relaxed vs. exact gradients on a small instance before scaling). Without this, the 123× improvement on the photonic metasensor cannot be confidently attributed to the joint-DP procedure rather than surrogate artifacts.

Authors: We agree that a controlled validation is required to attribute the gains confidently. We have added a new supplementary experiment on the radar beam-search POMDP (small enough for exact DP). Geometries optimized with the full relaxation hierarchy produce adaptive values within 2% of the exact-DP optimum, and the factor-of-2.8 improvement over the non-adaptive baseline is preserved. This result is now presented as Supplementary Figure S1 and referenced in the main text when discussing the photonic metasensor scaling. The comparison supports that the reported 123× MSE reduction is due to joint optimization rather than surrogate artifacts. revision: yes

Circularity Check

0 steps flagged

No circularity in joint-DP co-design derivation

full rationale

The paper formulates joint dynamic programming as an optimization over hardware geometry and Bellman-optimal policy, then invokes standard differentiable DP plus the envelope theorem to obtain exact outer gradients, followed by a relaxation hierarchy to scale the framework. These steps apply external mathematical results (Bellman optimality, envelope theorem) to the problem without any claimed prediction or result reducing by construction to a fitted parameter, self-citation chain, or definitional tautology. The assertion that joint optimization is the minimum principled procedure follows directly from the co-design formulation and its comparison to non-adaptive baselines, remaining self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 1 invented entities

The central claim rests on the differentiability and exactness of the DP gradient via the envelope theorem plus the validity of the relaxation hierarchy for large discrete topologies; no explicit free parameters are named in the abstract.

axioms (2)

domain assumption Differentiable dynamic programming with sharp Bellman maximum yields exact, bias-free outer gradients via the envelope theorem
Invoked to justify gradient computation for hardware geometry optimization.
domain assumption The relaxation hierarchy preserves solution quality when scaling from small POMDPs to 10^5-pixel photonic topologies
Required for the method to apply to the largest case study.

invented entities (1)

joint-DP framework no independent evidence
purpose: Single optimization over continuous hardware geometry and Bellman-optimal adaptive policy
New co-design procedure introduced in the paper; no independent evidence provided beyond the three case studies.

pith-pipeline@v0.9.0 · 5616 in / 1496 out tokens · 46269 ms · 2026-05-07T15:47:33.890826+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

127 extracted references · 28 canonical work pages

[1]

Joannopoulos, Steven G

John D. Joannopoulos, Steven G. Johnson, Joshua N. Winn, and Robert D. Meade. Photonic Crystals: Molding the Flow of Light. Princeton University Press, Princeton, NJ, 2nd edition, 2008

2008
[2]

Flat optics with designer metasurfaces.Nature Ma- terials, 13:139–150, 2014

Nanfang Yu and Federico Capasso. Flat optics with designer metasurfaces.Nature Ma- terials, 13:139–150, 2014. doi: 10.1038/nmat3839

work page doi:10.1038/nmat3839 2014
[3]

Piggott, Weiliang Jin, Jelena Vuckovic, and Ale- jandro W

Sean Molesky, Zin Lin, Alexander Y. Piggott, Weiliang Jin, Jelena Vuckovic, and Ale- jandro W. Rodriguez. Inverse design in nanophotonics.Nature Photonics, 12:659–670, 2018. 25

2018
[4]

Jensen and Ole Sigmund

Jakob S. Jensen and Ole Sigmund. Topology optimization for nano-photonics.Laser & Photonics Reviews, 5:308–321, 2011. doi: 10.1002/lpor.201000014

work page doi:10.1002/lpor.201000014 2011
[5]

Jiaqi Jiang, Mingkun Chen, and Jonathan A. Fan. Deep neural networks for the evaluation and design of photonic devices.Nature Reviews Materials, 6:679–700, 2021. doi: 10.1038/ s41578-020-00260-1

2021
[6]

Christiansen and Ole Sigmund

Rasmus E. Christiansen and Ole Sigmund. Inverse design in photonics by topology opti- mization: Tutorial.Journal of the Optical Society of America B, 38:496–509, 2021. doi: 10.1364/JOSAB.406048

work page doi:10.1364/josab.406048 2021
[7]

Zin Lin, Victor Liu, Rapha¨ el Pestourie, and Steven G. Johnson. Topology optimization of freeform large-area metasurfaces.Optics Express, 27:15765–15775, 2019

2019
[8]

Cummer, Johan Christensen, and Andrea Al` u

Steven A. Cummer, Johan Christensen, and Andrea Al` u. Controlling sound with acoustic metamaterials.Nature Reviews Materials, 1:16001, 2016. doi: 10.1038/natrevmats.2016.1

work page doi:10.1038/natrevmats.2016.1 2016
[9]

Price, Alberto Amo, Nathan Goldman, Mohammad Hafezi, Ling Lu, Mikael C

Tomoki Ozawa, Hannah M. Price, Alberto Amo, Nathan Goldman, Mohammad Hafezi, Ling Lu, Mikael C. Rechtsman, David Schuster, Jonathan Simon, Oded Zilberberg, and Iacopo Carusotto. Topological photonics.Reviews of Modern Physics, 91:015006, 2019. doi: 10.1103/RevModPhys.91.015006

work page doi:10.1103/revmodphys.91.015006 2019
[10]

Roy Sambles

Peter Vukusic and J. Roy Sambles. Photonic structures in biology.Nature, 424:852–855,
[11]

doi: 10.1038/nature01941

work page doi:10.1038/nature01941
[12]

Miller, Yiwen Yang, Steven G

Owen D. Miller, Yiwen Yang, Steven G. Johnson, et al. Fundamental limits to optical response in absorptive systems.Optics Express, 24:3329–3364, 2016

2016
[13]

Hyungki Shim, Zeyu Kuang, Zin Lin, and Owen D. Miller. Fundamental limits to multi- functional and tunable nanophotonic response.Nanophotonics, 13(12):2107–2116, 2024. doi: 10.1515/nanoph-2023-0630

work page doi:10.1515/nanoph-2023-0630 2024
[14]

End-to-end optimization of optics and image processing for achromatic extended depth of field and super-resolution imaging

Vincent Sitzmann, Steven Diamond, Yifan Peng, Xiong Dun, Stephen Boyd, Wolfgang Heidrich, Felix Heide, and Gordon Wetzstein. End-to-end optimization of optics and image processing for achromatic extended depth of field and super-resolution imaging. ACM Transactions on Graphics (SIGGRAPH), 2018

2018
[15]

Deep optics for monocular depth estimation and 3d object detection.ICCV, 2019

Julie Chang and Gordon Wetzstein. Deep optics for monocular depth estimation and 3d object detection.ICCV, 2019

2019
[16]

Metzler, Hayato Ikoma, Yifan Peng, and Gordon Wetzstein

Christopher A. Metzler, Hayato Ikoma, Yifan Peng, and Gordon Wetzstein. Deep optics for single-shot high-dynamic-range imaging.CVPR, 2020

2020
[17]

Neural nano-optics for high-quality thin lens imaging

Ethan Tseng, Shane Colburn, James Whitehead, Luocheng Huang, Seung-Hwan Baek, Arka Majumdar, and Felix Heide. Neural nano-optics for high-quality thin lens imaging. Nature Communications, 2021

2021
[18]

Kellman, Emrah Bostan, Nicole A

Michael R. Kellman, Emrah Bostan, Nicole A. Repina, and Laura Waller. Physics-based learned design: optimized coded-illumination for quantitative phase imaging.IEEE Trans- actions on Computational Imaging, 5:344–353, 2019

2019
[19]

FlexISP: A flexible camera image processing framework.ACM Transactions on Graphics (SIGGRAPH Asia), 2014

Felix Heide, Markus Steinberger, Yun-Ta Tsai, Mushfiqur Rouf, Dawid Pajak, Dikpal Reddy, Orazio Gallo, Jing Liu, Wolfgang Heidrich, Karen Egiazarian, Jan Kautz, and Kari Pulli. FlexISP: A flexible camera image processing framework.ACM Transactions on Graphics (SIGGRAPH Asia), 2014. 26

2014
[20]

DiffuserCam: lensless single-exposure 3d imaging.Optica, 2018

Nick Antipa, Grace Kuo, Reinhard Heckel, Ben Mildenhall, Emrah Bostan, Ren Ng, and Laura Waller. DiffuserCam: lensless single-exposure 3d imaging.Optica, 2018

2018
[21]

Li, Charles Roques-Carmes, Marin Soljaˇ ci´ c, Steven G

Gaurav Arya, William F. Li, Charles Roques-Carmes, Marin Soljaˇ ci´ c, Steven G. John- son, and Zin Lin. End-to-end optimization of metasurfaces for imaging with compressed sensing.ACS Photonics, 11(5):2077–2087, 2024. doi: 10.1021/acsphotonics.4c00259

work page doi:10.1021/acsphotonics.4c00259 2077
[22]

Zin Lin, Charles Roques-Carmes, Rapha¨ el Pestourie, Marin Soljaˇ ci´ c, Arka Majumdar, and Steven G. Johnson. End-to-end nanophotonic inverse design for imaging and polarimetry. Nanophotonics, 10:1177–1187, 2021

2021
[23]

Machine-learning reprogrammable metasurface imager.Nature Communications, 10:1082, 2019

Lianlin Li, Tie Jun Cui, Wei Ji, Shuo Liu, Jun Ding, Xiang Wan, Yun Bo Li, Menghua Jiang, Cheng-Wei Qiu, and Shuang Zhang. Machine-learning reprogrammable metasurface imager.Nature Communications, 10:1082, 2019. doi: 10.1038/s41467-019-09103-2

work page doi:10.1038/s41467-019-09103-2 2019
[24]

Littman, and Anthony R

Leslie Pack Kaelbling, Michael L. Littman, and Anthony R. Cassandra. Planning and acting in partially observable stochastic domains.Artificial Intelligence, 101:99–134, 1998

1998
[25]

Smallwood and Edward J

Richard D. Smallwood and Edward J. Sondik. The optimal control of partially observable Markov processes over a finite horizon.Operations Research, 21(5):1071–1088, 1973

1973
[26]

Sondik.The optimal control of partially observable Markov processes

Edward J. Sondik.The optimal control of partially observable Markov processes. PhD thesis, Stanford University, 1971

1971
[27]

Hero, III

Chris Kreucher, Keith Kastella, and Alfred O. Hero, III. Sensor management using an active sensing approach.Signal Processing, 85(3):607–624, 2005

2005
[28]

Hero and Douglas Cochran

Alfred O. Hero and Douglas Cochran. Sensor management: Past, present, and future. IEEE Sensors Journal, 11(12):3064–3075, 2011

2011
[29]

A. O. Hero, D. A. Casta˜ n´ on, D. Cochran, and K. Kastella, editors.Foundations and Applications of Sensor Management. Springer, 2008

2008
[30]

Near-optimal bayesian active learning with correlated and noisy tests

Yuxin Chen and Andreas Krause. Near-optimal bayesian active learning with correlated and noisy tests. InAISTATS, 2017

2017
[31]

Near-optimal nonmyopic value of information in graphical models

Andreas Krause and Carlos Guestrin. Near-optimal nonmyopic value of information in graphical models. InUAI, 2005

2005
[32]

D. V. Lindley. On a measure of the information provided by an experiment.Annals of Mathematical Statistics, 27(4):986–1005, 1956

1956
[33]

Xun Huan and Youssef M. Marzouk. Sequential Bayesian optimal experimental design via approximate dynamic programming.arXiv:1604.08320, 2016

work page Pith review arXiv 2016
[34]

Bayesian sequential optimal experimental design for nonlinear models using policy gradient reinforcement learning.arXiv:2110.15335, 2023

Wanggang Shen and Xun Huan. Bayesian sequential optimal experimental design for nonlinear models using policy gradient reinforcement learning.arXiv:2110.15335, 2023

work page arXiv 2023
[35]

Optimal experimental design for infinite-dimensional Bayesian inverse problems governed by PDEs: a review.Acta Numerica, 2021

Alen Alexanderian. Optimal experimental design for infinite-dimensional Bayesian inverse problems governed by PDEs: a review.Acta Numerica, 2021

2021
[36]

Bayesian experimental design: A review.Sta- tistical Science, 10(3):273–304, 1995

Kathryn Chaloner and Isabella Verdinelli. Bayesian experimental design: A review.Sta- tistical Science, 10(3):273–304, 1995

1995
[37]

Bell, Chris J

Kristine L. Bell, Chris J. Baker, Graeme E. Smith, Joel T. Johnson, and Muralidhar Rangaswamy. Cognitive radar framework for target detection and tracking.IEEE Journal of Selected Topics in Signal Processing, 2015. 27

2015
[38]

Cognitive radar: a way of the future.IEEE Signal Processing Magazine, 23:30–40, 2006

Simon Haykin. Cognitive radar: a way of the future.IEEE Signal Processing Magazine, 23:30–40, 2006

2006
[39]

Greco, Fulvio Gini, Pietro Stinco, and Kristine L

Maria S. Greco, Fulvio Gini, Pietro Stinco, and Kristine L. Bell. Cognitive radars: On the road to reality.IEEE Signal Processing Magazine, 2018

2018
[40]

POMDP-driven cognitive massive MIMO radar: Joint target detection-tracking in unknown disturbances

Imad Bouhou, Stefano Fortunati, Leila Gharsalli, and Alexandre Renaux. POMDP-driven cognitive massive MIMO radar: Joint target detection-tracking in unknown disturbances. IEEE Transactions on Radar Systems, 2025. arXiv:2410.17967

work page arXiv 2025
[41]

Quantum metrology.Physical Review Letters, 96:010401, 2006

Vittorio Giovannetti, Seth Lloyd, and Lorenzo Maccone. Quantum metrology.Physical Review Letters, 96:010401, 2006

2006
[42]

Braunstein and Carlton M

Samuel L. Braunstein and Carlton M. Caves. Statistical distance and the geometry of quantum states.Physical Review Letters, 72:3439–3443, 1994

1994
[43]

Matteo G. A. Paris. Quantum estimation for quantum technology.International Journal of Quantum Information, 7:125–137, 2009

2009
[44]

Carlton M. Caves. Quantum-mechanical noise in an interferometer.Physical Review D, 23:1693–1708, 1981

1981
[45]

Tunable transmon qubit as a flux sensor: joint device-protocol optimization.arXiv:2211.08344v4, 2024

Sergey Danilin, Nicholas Nugent, and Martin Weides. Tunable transmon qubit as a flux sensor: joint device-protocol optimization.arXiv:2211.08344v4, 2024

work page arXiv 2024
[46]

A. Yu. Kitaev. Quantum measurements and the Abelian stabilizer problem.arXiv:quant- ph/9511026, 1995

work page arXiv 1995
[47]

Braverman, A

B. Braverman, A. Kawasaki, E. Pedrozo-Pe˜ nafiel, S. Colombo, C. Shu, Z. Li, E. Mendez, M. Yamoah, L. Salvi, D. Akamatsu, et al. Near-unitary spin squeezing in 171yb.Physical Review Letters, 122:223203, 2019

2019
[48]

Gollub, Daniel L

Okan Yurduseven, Jonah N. Gollub, Daniel L. Marks, and David R. Smith. Printed aperiodic cavity for computational and microwave imaging.IEEE Microwave and Wireless Components Letters, 2017

2017
[49]

Imani, Jonah N

Mohammadreza F. Imani, Jonah N. Gollub, Okan Yurduseven, Aaron V. Diebold, Michael Boyarsky, Thomas Fromenteze, Laura Pulido-Mancera, Timothy Sleasman, and David R. Smith. Review of metasurface antennas for computational microwave imaging.IEEE Transactions on Antennas and Propagation, 2020

2020
[50]

Advances in photonic quantum sensing.Nature Photonics, 12:724–733, 2018

Stefano Pirandola, Bhaskar Roy Bardhan, Tobias Gehring, Christian Weedbrook, and Seth Lloyd. Advances in photonic quantum sensing.Nature Photonics, 12:724–733, 2018. doi: 10.1038/s41566-018-0301-6

work page doi:10.1038/s41566-018-0301-6 2018
[51]

C. L. Degen, F. Reinhard, and P. Cappellaro. Quantum sensing.Reviews of Modern Physics, 89:035002, 2017

2017
[52]

Princeton University Press, 1946

Harald Cram´ er.Mathematical Methods of Statistics. Princeton University Press, 1946

1946
[53]

C. R. Rao. Information and the accuracy attainable in the estimation of statistical pa- rameters.Bulletin of the Calcutta Mathematical Society, 37:81–91, 1945

1945
[54]

Envelope theorems for arbitrary choice sets.Econometrica, 70(2):583–601, 2002

Paul Milgrom and Ilya Segal. Envelope theorems for arbitrary choice sets.Econometrica, 70(2):583–601, 2002

2002
[55]

Benveniste and Jose A

Lawrence M. Benveniste and Jose A. Scheinkman. On the differentiability of the value function in dynamic models of economics.Econometrica, 47(3):727–732, 1979. 28

1979
[56]

Differentiable dynamic programming for structured prediction and attention

Arthur Mensch and Mathieu Blondel. Differentiable dynamic programming for structured prediction and attention. InProceedings of the 35th International Conference on Machine Learning (ICML), 2018

2018
[57]

Jimenez Rodriguez, Jacob Sacks, Byron Boots, and J

Brandon Amos, Ivan D. Jimenez Rodriguez, Jacob Sacks, Byron Boots, and J. Zico Kolter. Differentiable MPC for end-to-end planning and control. InAdvances in Neural Informa- tion Processing Systems (NeurIPS), 2018

2018
[58]

Generic methods for optimization-based modeling

Justin Domke. Generic methods for optimization-based modeling. InProceedings of the 15th International Conference on Artificial Intelligence and Statistics (AISTATS), 2012

2012
[59]

Efficient and modular implicit differentiation.Advances in Neural Information Processing Systems (NeurIPS), 2022

Mathieu Blondel, Quentin Berthet, Marco Cuturi, Roy Frostig, Stephan Hoyer, Felipe Llinares-L´ opez, Fabian Pedregosa, and Jean-Philippe Vert. Efficient and modular implicit differentiation.Advances in Neural Information Processing Systems (NeurIPS), 2022

2022
[60]

Van Trees and Kristine L

Harry L. Van Trees and Kristine L. Bell.Bayesian Bounds for Parameter Estimation and Nonlinear Filtering/Tracking. Wiley-IEEE Press, 2007

2007
[61]

Muravchik, and Arye Nehorai

Petr Tichavsk´ y, Carlos H. Muravchik, and Arye Nehorai. Posterior Cram´ er–Rao bounds for discrete-time nonlinear filtering.IEEE Transactions on Signal Processing, 46(5):1386– 1396, 1998

1998
[62]

Princeton University Press, 1957

Richard Bellman.Dynamic Programming. Princeton University Press, 1957

1957
[63]

Kellman, Emrah Bostan, Michael Chen, and Laura Waller

Michael R. Kellman, Emrah Bostan, Michael Chen, and Laura Waller. Data-driven design for Fourier ptychographic microscopy.Optica, 7:1075–1082, 2020

2020
[64]

Kim, Jaebum Park, Michael Chen, Laura Waller, and Roarke Horstmeyer

Alex Muthumbi, Amey Chaware, Kangwook Kim, Kevin C. Kim, Jaebum Park, Michael Chen, Laura Waller, and Roarke Horstmeyer. Learned sensing: jointly optimized micro- scope hardware for accurate image classification.Biomedical Optics Express, 10:6351– 6369, 2019

2019
[65]

Deep phase decoder: self-calibrating phase microscopy with an untrained deep neural network.Optica, 7:559–562, 2020

Emrah Bostan, Reinhard Heckel, Michael Chen, Michael Kellman, and Laura Waller. Deep phase decoder: self-calibrating phase microscopy with an untrained deep neural network.Optica, 7:559–562, 2020

2020
[66]

Kellman, Emrah Bostan, Nicole A

Michael R. Kellman, Emrah Bostan, Nicole A. Repina, and Laura Waller. Physics-based learned design: Optimized coded-illumination for quantitative phase imaging.IEEE Transactions on Computational Imaging, 5(3):344–353, 2019. doi: 10.1109/TCI.2019. 2905434

work page doi:10.1109/tci.2019 2019
[67]

John M. Danskin. The theory of max-min, with applications.SIAM Journal on Applied Mathematics, 14:641–664, 1966

1966
[68]

Ronald A. Howard. Information value theory.IEEE Transactions on Systems Science and Cybernetics, 2(1):22–26, 1966

1966
[69]

Harvard Busi- ness School, 1961

Howard Raiffa and Robert Schlaifer.Applied Statistical Decision Theory. Harvard Busi- ness School, 1961

1961
[70]

Rodriguez

Pengning Chao, Benjamin Strekha, others, Sean Molesky, and Alejandro W. Rodriguez. Maximum electromagnetic local density of states via material structuring: T-matrix bounds.Nature Reviews Physics, 2022

2022
[71]

Venkataram, Weiliang Jin, and Alejandro W

Sean Molesky, Prashanth S. Venkataram, Weiliang Jin, and Alejandro W. Rodriguez. Fundamental limits to radiative heat transfer: theory.Physical Review B, 2020. 29

2020
[72]

Rodriguez

Sean Molesky, Pengning Chao, and Alejandro W. Rodriguez. Hierarchical mean-field t-matrix bounds on photonic scattering.Physical Review Research, 2020

2020
[73]

Rodriguez

Sean Molesky, Pengning Chao, Weiliang Jin, and Alejandro W. Rodriguez. T-operator limits on optical communication: metaoptics, computation, and input-output transforma- tions.Physical Review Research, 2022

2022
[74]

Zeyu Kuang and Owen D. Miller. Computational bounds to light-matter interactions via local conservation laws.Physical Review Letters, 2020

2020
[75]

Rodriguez

Pengning Chao, Benjamin Strekha, Rodrick Kuate Defo, Sean Molesky, and Alejandro W. Rodriguez. Physical limits in electromagnetism.Nature Reviews Physics, 4:543–559, 2022. doi: 10.1038/s42254-022-00468-w

work page doi:10.1038/s42254-022-00468-w 2022
[76]

Upper bounds on absorption and scattering.New Journal of Physics, 2020

Mats Gustafsson, Kurt Schab, Lukas Jelinek, and Miloslav Capek. Upper bounds on absorption and scattering.New Journal of Physics, 2020

2020
[77]

Radiation efficiency limits in direct antenna modulation transmitters.IEEE Antennas and Wireless Propagation Letters, 19 (11):1911–1915, 2020

Kurt Schab, Lukas Jelinek, and Miloslav Capek. Radiation efficiency limits in direct antenna modulation transmitters.IEEE Antennas and Wireless Propagation Letters, 19 (11):1911–1915, 2020. doi: 10.1109/LAWP.2020.3020657

work page doi:10.1109/lawp.2020.3020657 1911
[78]

Rodriguez

Sean Molesky, Pengning Chao, Weiliang Jin, and Alejandro W. Rodriguez. Global T operator bounds on electromagnetic scattering: Upper bounds on far-field cross sections. Physical Review Research, 2:033172, 2020. doi: 10.1103/PhysRevResearch.2.033172

work page doi:10.1103/physrevresearch.2.033172 2020
[79]

Venkataram, and Alejandro W

Sean Molesky, Weiliang Jin, Prashanth S. Venkataram, and Alejandro W. Rodriguez. T operator bounds on angle-integrated absorption and thermal radiation for arbitrary ob- jects.Physical Review Letters, 123:257401, 2019. doi: 10.1103/PhysRevLett.123.257401

work page doi:10.1103/physrevlett.123.257401 2019
[80]

Venkataram, Sean Molesky, Weiliang Jin, and Alejandro W

Prashanth S. Venkataram, Sean Molesky, Weiliang Jin, and Alejandro W. Rodriguez. Fundamental limits to radiative heat transfer: the limited role of nanostructuring in the near-field.Physical Review Letters, 2020

2020

Showing first 80 references.