Pith. sign in

REVIEW 2 major objections 5 minor 43 references

Measurement-Access Risk Frontiers for Autonomous Scientific Control

T0 review · 2 major / 5 minor · reviewed 2026-07-11 · grok-4.5

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Autonomous science cannot collapse decision uncertainty by computation alone; residual risk is set by which physical records a platform exposes before it acts.

desk verdict Solid architectural packaging of a classical Bayes projection fact into a pre-deployment measurement-access audit for autonomous labs; math is correct, novelty is framing not foundations. read the letter →

arxiv 2607.05696 v1 pith:7RB25DL5 submitted 2026-07-06 math-ph math.MPphysics.chem-phphysics.data-an

Bo Peng This is my paper · ORCID
classification math-phmath.MPphysics.chem-phphysics.data-an
keywords physicallyaccessibledecision-makingmeasurement-accessriskfrontierno-free-autonomyautonomousscienceBayesfloormonitoredfeedbacksensoraudithiddenregimes
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper argues that scaling autonomous science is limited not only by algorithms, compute, or data volume, but by which physical records a platform can actually generate before it chooses an action. The authors formulate physically accessible decision-making (PADM) and a measurement-access risk frontier: the Bayes-optimal target risk minimized over records realizable under cost, bandwidth, latency, disturbance, memory, and actuation constraints. From this they derive a no-free-autonomy limit: an optimal controller cannot remove target components that never enter its record, so residual risk is an access floor rather than an optimizer failure. Closing that gap requires expanded sensing, auditing, tolerated disturbance, slower or staged operation, or restricted deployment. Concrete checks include monitored feedback with a hidden switching force, Gaussian and hidden-regime benchmarks, cost-aware and thermodynamic channel selection, and a chemistry-aware ranking audit on a 1000-target stress panel.

What carries the argument

The measurement-access risk frontier R*_auto(Λ), carried by the autonomous-record risk floor (Theorem 1): for a scalar target T and pre-action record Y_A, under quadratic loss any square-integrable action a(Y_A) has risk at least E[(T − E[T|Y_A])²], with equality at the Bayes action. The projection identity converts residual risk into a diagnostic of missing physical access, and complementary-channel value equals how much the added record changes that target projection.

What would settle it

On a fixed real autonomous-lab target, add a candidate audit channel that should change the target projection; if held-out decision risk does not fall relative to the autonomous record alone while computational limits are controlled, the access floor is not the binding limit the audit claims to identify.

Watch

Extended reading notes

Core claim

The central claim is that autonomous scientific control is bounded by a measurement-access risk frontier: the best Bayes risk attainable over physically realizable pre-action records under platform constraints. Under quadratic loss, no controller using only the autonomous record can beat the residual variance left after projecting the target onto that record. That residual is therefore a physical access floor, not a computational defect, and the gap to an oracle closes only by expanding access, adding audit channels, slowing the loop, or restricting the operating domain.

Load-bearing premise

The paper treats the known-model Bayes residual under a chosen target and an idealized oracle as the right pre-deployment diagnostic, even though real systems also face model error, finite samples, and approximate inference.

Editorial extensions

If this is right

  • Pre-deployment audits should score candidate sensors by target-specific risk recovery, not by generic information content.
  • A platform that acts before exposing a target-relevant mode retains irreducible residual risk no matter how strong the learning algorithm.
  • Closing an automation gap requires expanded sensing, audit channels, slower staged operation, or a restricted autonomous domain.
  • Redundant records leave the risk floor unchanged; only complementary access that moves the Bayes projection of the target reduces risk.
  • Cost, bandwidth, latency, and back-action enter as constraints on the feasible record set, so instrument and thermodynamic limits become part of the decision bound.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Self-driving laboratories may need formal measurement-sufficiency checks before full closed-loop deployment, analogous to safety cases for acting under partial observability.
  • Many unexpected failures in robotic chemistry and materials loops could reclassify as missing channels rather than weak acquisition functions or under-trained policies.
  • Quantum feedback already distinguishes monitored channels from unmonitored modes; the same target-risk recovery criterion could rank channels under back-action cost.
  • If access gaps routinely dominate learning error in pilots, investment should shift from larger models toward instrument architecture and pre-action audit sensors.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

2 major / 5 minor

Summary. The manuscript formulates physically accessible decision-making (PADM) and a measurement-access risk frontier R*_auto(\Lambda): the Bayes-optimal target risk minimized over records realizable under cost, bandwidth, latency, disturbance, memory and actuation constraints. Its central claim is a no-free-autonomy limit: under quadratic loss, any square-integrable action based only on the pre-action autonomous record Y_A cannot beat residual variance of the target T given Y_A (Theorem 1), so an optimal controller cannot remove target components absent from its record (Corollary 1). Closing the gap requires expanded access, auditing, tolerated disturbance, slower operation or restricted deployment. The claim is illustrated by monitored feedback with a hidden switching force, a chemistry-aware candidate-ranking audit with a 1000-target stress panel, Gaussian sensing, hidden-regime decisions and cost-aware/thermodynamic channel selection, plus an operational audit algorithm.

Significance. If the framing holds as a pre-deployment diagnostic, it supplies a clean architecture-level language for when autonomous scientific platforms are measurement-insufficient rather than merely under-optimized. The paper is explicit that Theorem 1 is the classical L2 projection identity and that the contribution is the frontier over constrained measurement architectures, residual oracle gaps and the audit workflow. Strengths include closed-form residual-risk formulas and equality cases for the Gaussian and hidden-regime benchmarks, a deterministic 1000-target chemistry stress panel with explicit recovery counts, monitored-feedback residual risk checked against the projection identity, and a stated retrospective CSV audit protocol with a public CAMEO/NIST shadow-mode check. These make the access floor falsifiable in the known-model setting and useful as a preflight check alongside experimental design, POMDPs and RL rather than as a replacement for them.

major comments (2)
  1. Discussion and Methods: the paper correctly states that finite-sample error, misspecification, learned representations, approximate inference and optimizer failure can only raise realized risk above the access floor, but the operational claim that PADM identifies residual oracle gaps before deployment still depends on estimating \Gamma_or and channel gains G_j / A_j from pilot data. The manuscript should state more explicitly under what pilot-data conditions those estimators remain informative (e.g., blocked splits, negative controls, distribution shift), and what decision the audit should return when estimation noise or misspecification is comparable to the reported gap, so that the diagnostic is not over-read as binding whenever computational errors dominate.
  2. Section 1.1, Eq. (1) and the measurement-sufficiency condition: R*_auto(\Lambda) is defined as an infimum over the feasible autonomous record set Y_auto_acc(\Lambda), yet the main demonstrations fix particular architectures (displacement-only vs cue; descriptor-only vs chemistry audit) rather than characterizing or approximating that set under concrete \Lambda. A short constructive statement of how Y_auto_acc is delimited for at least one platform class (e.g., monitored feedback with bandwidth/latency, or the chemistry audit with pre-action timing) would make the frontier operational rather than primarily conceptual.
minor comments (5)
  1. Figure 2c: the sufficiency contour \Delta R=0.10 is illustrative; state in the caption or Methods that it is a chosen threshold, not a universal criterion, and how it maps to the task-specific \epsilon in the sufficiency definition.
  2. Figure 3 and the chemistry panel: emphasize earlier in the main text (not only in the caption and Discussion) that this is a fixed-seed adversarial ranking stress test, not experimental chemical validation or a competitive generator benchmark.
  3. Algorithm 1: the recovered-oracle-gap fraction A_j is undefined when R(Y_A)=R(Y_or); the skip rule is stated, but a one-line note that A_j is then omitted (or set to 1 by convention) would avoid implementation ambiguity.
  4. Notation: Y_auto_acc(\Lambda), P_adm(Y), \Gamma_or and R_dyn(Y) are introduced densely in Section 1.1–1.2; a short symbol table or consistent first-use expansion would help readers outside mathematical physics.
  5. Code availability: the manuscript states code will be public upon publication and is available to editors/reviewers on request; for reproducibility of the 1000-target panel and unit tests, a stable archive identifier at acceptance would strengthen the claim.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the load-bearing claim is the classical L2 projection identity applied to constrained pre-action records, with synthetic checks against closed forms rather than fitted-as-prediction loops.

full rationale

Theorem 1 is the standard orthogonal-projection / conditional-expectation identity under quadratic loss (E[(T-a(Y_A))^2] >= residual variance of T given Y_A). The paper states this explicitly as classical and uses it as the proof engine for a measurement-access framing, not as a novel derivation that smuggles the conclusion from fitted inputs. Corollary 1 and the risk-frontier definition follow by restricting the feasible pre-action record family; they are not statistically forced predictions. Benchmarks (linear-Gaussian precision update, hidden-regime closed forms R_A, R_A+H, R_oracle, monitored-feedback residual E[(z_t-q_t)^2], cost-aware convex solution) are checked against analytic formulas or Monte Carlo of the same models. The chemistry 1000-target panel is an adversarial stress test that selects the inserted artifact 1000/1000 by construction under the descriptor-only rule—the paper labels it as such and as a computational audit benchmark, not empirical discovery. There is no load-bearing self-citation chain, no uniqueness theorem imported from the same author, and no parameter fitted to data then re-presented as an independent prediction. Mild definitional packaging (residual risk named as access floor / no-free-autonomy) is ordinary theory framing, not circular reduction of a claimed first-principles result to its inputs.

Assumptions & free parameters 4 free parameters · 5 assumptions · 3 invented entities

The load-bearing math is standard L2 Bayes decision theory plus classical stochastic models. The paper's own additions are definitional objects (feasible autonomous record sets under Lambda, measurement-access frontier, PADM audit scores) and illustrative parameters in synthetic examples. No new physical force or particle is postulated; the hidden force, latent regime, and chemistry artifact are model components chosen to expose the access floor.

free parameters (4)
  • sufficiency tolerance epsilon = task-specific
    Task-specific threshold for declaring measurement-sufficiency (Gamma_or <= epsilon); chosen by the user rather than derived.
  • monitored-feedback defaults (Delta t, k, D, f, gamma, sigma_r, m) = Delta t=0.02, k=1.5, D=0.2, f=1, gamma=0.08, sigma_r=0.85, m=8
    Simulation parameters for the oscillator example and accessibility map; illustrative, not fitted to experiment.
  • cue reliability p and target scale beta in hidden-regime benchmark = beta=1; p in [0.5,1]
    Analytic/Monte Carlo benchmark knobs controlling complementary-channel value; chosen for illustration.
  • directional sensing costs in cost-aware Gaussian benchmark
    Relative costs that make the hidden direction expensive under autonomous access; set to demonstrate cost-aware channel selection.
assumptions (5)
  • standard math Bayes decision theory: optimal action minimizes expected loss given the available observation; under quadratic loss the optimum is the conditional expectation.
    Used throughout the frontier definition and Theorem 1.
  • standard math L2 orthogonal projection identity: residual risk after conditioning on a record is the unexplained variance of the target.
    Proof engine of Theorem 1 and the channel-value formula Delta_T.
  • domain assumption Feasible autonomous records are those realizable under physical constraint vector Lambda (cost, bandwidth, latency, disturbance, memory, actuation).
    Defines Y_auto_acc(Lambda) and therefore the frontier R*_auto(Lambda).
  • domain assumption Known-model setting: the operating distribution of S, measurement channels, and loss are treated as known when stating the access bound; computational errors only raise realized risk.
    Stated in Results 1.3 and Discussion as the scope of the bound.
  • ad hoc to paper Oracle access is an ideal reference exposing target-relevant variables, not a physically available controller.
    Needed to define Gamma_or and measurement-sufficiency; explicitly model-relative.
invented entities (3)
  • physically accessible decision-making (PADM)
    purpose: Name the architecture-level audit of whether pre-action records expose target-relevant variables under physical constraints.
    Framework label for the frontier-plus-audit workflow; not an independent physical object.
  • measurement-access risk frontier R*_auto(Lambda)
    purpose: Define the best Bayes risk over records realizable under constraint vector Lambda.
    Central constructed object of the paper; reduces to ordinary Bayes risk once the feasible record set is fixed.
  • target-specific audit gain G_j / recovered oracle-gap fraction A_j
    purpose: Score candidate expanded channels by reduction in target residual risk rather than generic information.
    Operational scores in Algorithm 1; derived from the projection identity.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Measurement-Access Risk Frontiers for Autonomous Scientific Control." pith.science (2026). https://pith.science/paper/7RB25DL5

@misc{pith2026260705696,
  author       = {Pith},
  title        = {Pith review of: Measurement-Access Risk Frontiers for Autonomous Scientific Control},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/7RB25DL5}},
  note         = {Machine review of arXiv:2607.05696}
}
read the original abstract

Rapidly scaling autonomous science is limited not only by algorithms, compute or data volume, but by which physical records a platform exposes before action. We formulate physically accessible decision-making (PADM) and a measurement-access risk frontier: the Bayes-optimal target risk minimized over records realizable under cost, bandwidth, latency, disturbance, memory and actuation constraints. The frontier gives a no-free-autonomy limit: automation cannot collapse decision uncertainty by computation alone; an optimal controller cannot remove target components absent from its record, and closing that gap requires expanded access, auditing, tolerated disturbance, slower operation or restricted deployment. In monitored feedback, displacement-only control remains exposed to a hidden switching force, whereas a finite-bandwidth cue recovers part of the missing projection before action. A chemistry-aware candidate-ranking audit with a 1000-target stress panel, Gaussian sensing, hidden-regime decisions and cost-aware/thermodynamic channel selection provide reproducible checks. PADM identifies target-specific audit value and residual oracle gaps before deployment.

Figures

Figures reproduced from arXiv: 2607.05696 by the authors.

Figure 1
Figure 1. PADM procedure for auditing measurement access in autonomous scien￾tific control. The numbered path separates the procedure into four parts. (1) The sens￾ing/measurement block defines the autonomous record available before action, including spec￾tra, images, scalar readouts, model predictions and experiment logs. (2) The controller uses only this record to infer the experimental state and plan the next intervention.… view at source ↗
Figure 2
Figure 2. Open-system feedback is limited by finite-bandwidth access to hidden modes. A feedback controller observes a displacement record while a hidden environmental force switches sign. (a) Controlled oscillator, displacement record, hidden force and finite-bandwidth auxiliary cue. The lower trace illustrates that the cue y(t) enters the controller record only after sampling or filtering, rather than as instantaneous acces… view at source ↗
Figure 3
Figure 3. Chemistry-aware candidate-ranking audit for known-target recovery in de novo molecular-design workflows. (a) A deterministic template generator produces molecular strings and proposal history for the autonomous record YA; a chemistry audit record Yj adds RDKit-derived sanitization, single-component connectivity, valence/plausibility flags, forbidden￾substructure checks, QED/synthetic-accessibility proxies and a pre-… view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Minimal hidden-regime theorem benchmark for the decision risk floor. A visible record can be measured perfectly while still being insufficient for autonomous control if the target action depends on a hidden regime. (a) Two hidden regimes have the same visible record x,…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references · 43 canonical work pages

  1. [1]

    King, Jem Rowland, Stephen G

    Ross D. King, Jem Rowland, Stephen G. Oliver, Michael Young, Wayne Aubrey, Emma Byrne, Maria Liakata, Magdalena Markham, Pinar Pir, Larisa N. Soldatova, Andrew Sparkes, Ken- neth E. Whelan, and Amanda Clare. The automation of science.Science, 324(5923):85–89, 2009

    work page 2009

  2. [2]

    Roch, and Al´ an Aspuru-Guzik

    Florian H¨ ase, Lo¨ ıc M. Roch, and Al´ an Aspuru-Guzik. Next-generation experimentation with self-driving laboratories.Trends in Chemistry, 1(3):282–291, 2019. doi: 10.1016/j.trechm. 2019.02.007

    work page doi:10.1016/j.trechm 2019

  3. [3]

    Gilad Kusne, Jason Hattrick-Simpers, Keith A

    Eric Stach, Brian DeCost, A. Gilad Kusne, Jason Hattrick-Simpers, Keith A. Brown, Karina G. Reyes, Joshua Schrier, Simon Billinge, Tonio Buonassisi, Ian Foster, Carla P. Gomes, John M. Gregoire, Apurva Mehta, Joseph Montoya, Elsa Olivetti, Chanyeol Park, Eli Rotenberg, Semion K. Saikin, Stephanie Smullin, Valentin Stanev, and Benji Maruyama. Autonomous ex...

    work page doi:10.1016/j.matt.2021.06.036 2021

  4. [4]

    The rise of self-driving labs in chemical and materials sciences.Nature Synthesis, 2:483–492, 2023

    Milad Abolhasani and Eugenia Kumacheva. The rise of self-driving labs in chemical and materials sciences.Nature Synthesis, 2:483–492, 2023. doi: 10.1038/s44160-022-00231-0. 13

    work page doi:10.1038/s44160-022-00231-0 2023

  5. [5]

    Schmid, Sterling G

    Gary Tom, Stefan P. Schmid, Sterling G. Baird, Yang Cao, Kourosh Darvish, Han Hao, Stanley Lo, Sergio Pablo-Garc´ ıa, Ella M. Rajaonson, Marta Skreta, Naruki Yoshikawa, Samantha Corapi, Gun Deniz Akkoc, Felix Strieth-Kalthoff, Martin Seifrid, and Al´ an Aspuru-Guzik. Self-driving laboratories for chemistry and materials science.Chemical Reviews, 124(16):9...

    work page doi:10.1021/acs.chemrev.4c00055 2024

  6. [6]

    Maffettone, Vladimir V

    Benjamin Burger, Phillip M. Maffettone, Vladimir V. Gusev, Catherine M. Aitchison, Yang Bai, Xiaoyan Wang, Xiaobo Li, Ben M. Alston, Buyi Li, Rob Clowes, Nicola Rankin, Brandon Harris, Reiner S. Sprick, and Andrew I. Cooper. A mobile robotic chemist.Nature, 583: 237–241, 2020

    work page 2020

  7. [7]

    MacLeod, Frederick G

    Benjamin P. MacLeod, Frederick G. L. Parlane, Timothy D. Morrissey, Florian H¨ ase, Lo¨ ıc M. Roch, Kevan E. Dettelbach, Renato Moreira, Lars P. E. Yunker, Mary B. Rooney, Jason R. Deeth, Vitaliy Lai, Gillian J. Ng, Helen Situ, Ruo Zhang, Matthew S. Elliott, Thomas H. Haley, David J. Dvorak, Al´ an Aspuru-Guzik, Jason E. Hein, and Curtis P. Berlinguette. ...

    work page 2020

  8. [8]

    Gilad Kusne, Hongyu Yu, Congyu Wu, Hongyi Zhang, Jason Hattrick-Simpers, Brian DeCost, Sayantan Sarker, Corey Oses, Cormac Toher, Stefano Curtarolo, Albert V

    A. Gilad Kusne, Hongyu Yu, Congyu Wu, Hongyi Zhang, Jason Hattrick-Simpers, Brian DeCost, Sayantan Sarker, Corey Oses, Cormac Toher, Stefano Curtarolo, Albert V. Davydov, Ritesh Agarwal, Leonid A. Bendersky, Mo Li, Apurva Mehta, and Ichiro Takeuchi. On-the-fly closed-loop materials discovery via bayesian active learning.Nature Communications, 11:5966, 2020

    work page 2020

Show all 43 references
  1. [9]

    Szymanski et al

    Nathan J. Szymanski et al. An autonomous laboratory for the accelerated synthesis of inorganic materials.Nature, 624:86–91, 2023. doi: 10.1038/s41586-023-06734-w

    2023 doi

  2. [10]

    Schoenholz, Muratahan Aykol, Gregor Cheon, Ekin D

    Amil Merchant, Simon Batzner, Samuel S. Schoenholz, Muratahan Aykol, Gregor Cheon, Ekin D. Cubuk, et al. Scaling deep learning for materials discovery.Nature, 624:80–85, 2023

    2023

  3. [11]

    Real-time quantum feedback prepares and stabilizes photon number states.Nature, 477:73–77, 2011

    Cl´ ement Sayrin, Igor Dotsenko, Xingxing Zhou, Bruno Peaudecerf, Th´ eo Rybarczyk, S´ ebastien Gleyzes, Pierre Rouchon, Mazyar Mirrahimi, Hadis Amini, Michel Brune, Jean-Michel Rai- mond, and Serge Haroche. Real-time quantum feedback prepares and stabilizes photon number stat...

    2011

  4. [12]

    K. W. Murch, S. J. Weber, C. Macklin, and I. Siddiqi. Observing single quantum trajectories of a superconducting quantum bit.Nature, 502:211–214, 2013

    2013

  5. [13]

    Minev, Shantanu O

    Zlatko K. Minev, Shantanu O. Mundhada, Shyam Shankar, Philip Reinhold, R. Guti´ errez- J´ auregui, R. J. Schoelkopf, Mazyar Mirrahimi, H. J. Carmichael, and M. H. Devoret. To catch and reverse a quantum jump mid-flight.Nature, 570:200–204, 2019

    2019

  6. [14]

    Wiley, 1950

    Abraham Wald.Statistical Decision Functions. Wiley, 1950

    1950

  7. [15]

    Savage.The Foundations of Statistics

    Leonard J. Savage.The Foundations of Statistics. Wiley, 1954

    1954

  8. [16]

    DeGroot.Optimal Statistical Decisions

    Morris H. DeGroot.Optimal Statistical Decisions. McGraw-Hill, 1970

    1970

  9. [17]

    Berger.Statistical Decision Theory and Bayesian Analysis

    James O. Berger.Statistical Decision Theory and Bayesian Analysis. Springer, 1985

    1985

  10. [18]

    Harvard University Press, 1961

    Howard Raiffa and Robert Schlaifer.Applied Statistical Decision Theory. Harvard University Press, 1961. 14

    1961

  11. [19]

    Comparison of experiments

    David Blackwell. Comparison of experiments. InProceedings of the Second Berkeley Sympo- sium on Mathematical Statistics and Probability, pages 93–102. University of California Press, 1951

    1951

  12. [20]

    Equivalent comparisons of experiments.The Annals of Mathematical Statis- tics, 24(2):265–272, 1953

    David Blackwell. Equivalent comparisons of experiments.The Annals of Mathematical Statis- tics, 24(2):265–272, 1953

    1953

  13. [21]

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

    1956

  14. [22]

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

    1966

  15. [23]

    Jones, Matthias Schonlau, and William J

    Donald R. Jones, Matthias Schonlau, and William J. Welch. Efficient global optimization of expensive black-box functions.Journal of Global Optimization, 13:455–492, 1998

    1998

  16. [24]

    Adams, and Nando de Freitas

    Bobak Shahriari, Kevin Swersky, Ziyu Wang, Ryan P. Adams, and Nando de Freitas. Taking the human out of the loop: A review of bayesian optimization.Proceedings of the IEEE, 104 (1):148–175, 2016

    2016

  17. [25]

    ˚Astr¨ om

    Karl J. ˚Astr¨ om. Optimal control of markov processes with incomplete state information. Journal of Mathematical Analysis and Applications, 10(1):174–205, 1965

    1965

  18. [26]

    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

  19. [27]

    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(1–2):99–134, 1998

    1998

  20. [28]

    Bertsekas.Dynamic Programming and Optimal Control

    Dimitri P. Bertsekas.Dynamic Programming and Optimal Control. Athena Scientific, 2012

    2012

  21. [29]

    American Mathe- matical Society and Oxford University Press, 2000

    Shun-ichi Amari and Hiroshi Nagaoka.Methods of Information Geometry. American Mathe- matical Society and Oxford University Press, 2000

    2000

  22. [30]

    Springer, 2016

    Shun-ichi Amari.Information Geometry and Its Applications. Springer, 2016

    2016

  23. [31]

    Stochastic thermodynamics, fluctuation theorems and molecular machines.Re- ports on Progress in Physics, 75(12):126001, 2012

    Udo Seifert. Stochastic thermodynamics, fluctuation theorems and molecular machines.Re- ports on Progress in Physics, 75(12):126001, 2012

    2012

  24. [32]

    Generalized jarzynski equality under nonequilibrium feedback control.Physical Review Letters, 104:090602, 2010

    Takahiro Sagawa and Masahito Ueda. Generalized jarzynski equality under nonequilibrium feedback control.Physical Review Letters, 104:090602, 2010

    2010

  25. [33]

    Horowitz and Massimiliano Esposito

    Jordan M. Horowitz and Massimiliano Esposito. Thermodynamics with continuous information flow.Physical Review X, 4:031015, 2014

    2014

  26. [34]

    Juan M. R. Parrondo, Jordan M. Horowitz, and Takahiro Sagawa. Thermodynamics of infor- mation.Nature Physics, 11:131–139, 2015

    2015

  27. [35]

    Irreversibility and heat generation in the computing process.IBM Journal of Research and Development, 5(3):183–191, 1961

    Rolf Landauer. Irreversibility and heat generation in the computing process.IBM Journal of Research and Development, 5(3):183–191, 1961

    1961

  28. [36]

    Charles H. Bennett. The thermodynamics of computation–a review.International Journal of Theoretical Physics, 21:905–940, 1982. 15

    1982

  29. [37]

    Wiseman and Gerard J

    Howard M. Wiseman and Gerard J. Milburn.Quantum Measurement and Control. Cambridge University Press, 2009

    2009

  30. [38]

    Kurt Jacobs and Daniel A. Steck. A straightforward introduction to continuous quantum measurement.Contemporary Physics, 47(5):279–303, 2006

    2006

  31. [39]

    Luc Bouten, Ramon Van Handel, and Matthew R. James. An introduction to quantum filter- ing.SIAM Journal on Control and Optimization, 46(6):2199–2241, 2007

    2007

  32. [40]

    Volk and Milad Abolhasani

    Amanda A. Volk and Milad Abolhasani. Performance metrics to unleash the power of self- driving labs in chemistry and materials science.Nature Communications, 15:1378, 2024. doi: 10.1038/s41467-024-45569-5

    2024 doi

  33. [41]

    Griesbach, Rui Zhang, et al

    Shi Xuan Leong, Caleb E. Griesbach, Rui Zhang, et al. Steering towards safe self-driving laboratories.Nature Reviews Chemistry, 9:707–722, 2025. doi: 10.1038/s41570-025-00747-x

    2025 doi

  34. [42]

    Oxford University Press, 2002

    Heinz-Peter Breuer and Francesco Petruccione.The Theory of Open Quantum Systems. Oxford University Press, 2002

    2002

  35. [43]

    Rdkit: Open-source cheminformatics.https://www.rdkit.org, 2025

    Greg Landrum et al. Rdkit: Open-source cheminformatics.https://www.rdkit.org, 2025. Accessed for deterministic molecule parsing, descriptors and fingerprints. 16

    2025

Pith tools

Reviewed July 11, 2026 · model on record in the stance chip above.