arxiv: 2604.08130 · v1 · submitted 2026-04-09 · 📡 eess.SY · cs.SY

Recognition: 3 theorem links

· Lean Theorem

Cognitive Flexibility as a Latent Structural Operator for Bayesian State Estimation

Thanana Nuchkrua , Sudchai Boonto , Xiaoqi Liu

Authors on Pith no claims yet

Pith reviewed 2026-05-10 17:52 UTC · model grok-4.3

classification 📡 eess.SY cs.SY

keywords cognitive flexibilitybayesian filteringlatent structure selectionstructural mismatchinnovation scorestate estimationbelief recursionadaptive filtering

0 comments

The pith

Cognitive Flexibility acts as an online operator to select latent structures in Bayesian state estimation while preserving the filtering recursion.

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

The paper introduces Cognitive Flexibility as a representation-level operator for selecting latent structures in deep stochastic state-space models. It does this online using an innovation-based predictive score while keeping the Bayesian filtering recursion intact. Structural mismatch is defined as irreducible predictive inconsistency when the structure is fixed. The belief-structure recursion is proven well-posed with a structural descent property and finite switching, falling back to standard filtering when the model matches reality. Experiments confirm better accuracy under mismatch without harming performance in correct cases.

Core claim

Cognitive Flexibility is introduced as a representation-level operator that selects latent structures online via an innovation-based predictive score, while preserving the Bayesian filtering recursion. Structural mismatch is formalized as irreducible predictive inconsistency under fixed structure. The resulting belief-structure recursion is shown to be well posed, to exhibit a structural descent property, and to admit finite switching, with reduction to standard Bayesian filtering under correct specification.

What carries the argument

Cognitive Flexibility (CF) as a representation-level operator that selects latent structures online via an innovation-based predictive score while preserving the Bayesian filtering recursion.

Load-bearing premise

The innovation-based predictive score can detect and resolve structural mismatches without creating new inconsistencies or violating the Bayesian recursion.

What would settle it

A demonstration that the operator produces infinite structure switches or that the updated beliefs fail to satisfy the standard Bayesian update equations under mismatch.

Figures

Figures reproduced from arXiv: 2604.08130 by Sudchai Boonto, Thanana Nuchkrua, Xiaoqi Liu.

**Figure 2.** Figure 2: Experiment 4.1. Top: True state zt and estimates from Fixed LIN, Fixed NL, IMM, and CF, initialised at s0 = slin. Bottom: Structure sequence st (LIN= 0, NL= 1) and innovation scores ΦLIN, ΦNL; CF commits to snl at t ≈ 5 and produces no further switches. t 0 100 200 300 400 La t e n t s t a t e −20 0 20 t = 𝜏 Observation shift t 0 100 200 300 400 s t QUAD = 0 SAT = 1 t = 𝜏 Observation shift Φ ( 𝔅t ,s ) = − … view at source ↗

**Figure 3.** Figure 3: Experiment 4.2: Top: True state zt and estimates from fixed QUAD, fixed SAT, and CF. The dashed line marks the change time τ . Bottom: Selected structure st (QUAD= 0, SAT= 1) and scores ΦQUAD, ΦSAT. t 0 100 200 300 400 La t e n t s t a t e −20 −10 0 10 20 t 0 100 200 300 400 s t QUAD = 0 SAT = 1 Φ ( 𝔅t ,s ) = − lo g ℓ 𝜃,s 0 100 200 300 true zt fixed SAT fixed QUAD CF CF ΦQUAD ΦSAT [PITH_FULL_IMAGE:figures… view at source ↗

**Figure 4.** Figure 4: Experiment 4.3 (negative control, no observation shift). Top: True state zt and estimates from Fixed-QUAD, Fixed-SAT, and CF (proposed); CF overlaps with Fixed-QUAD throughout. Bottom: Structure sequence st (QUAD= 0, SAT= 1) and innovation scores ΦQUAD, ΦSAT; the selected structure remains at QUAD for all t and ΦQUAD stays below ΦSAT throughout. while under ssat, yt ∼ N tanh z 2 t 20 , σˆ 2 v . (31) … view at source ↗

**Figure 5.** Figure 5: Experiment 4.4 (Z = R 2 , T = 200, N = 1000 particles, s0 = slin). Panels 1–2: Latent state estimates zt,1 and zt,2; CF recovers the accuracy of Fixed NL in both dimensions after a single structural switch, while Fixed LIN remains persistently biased. Panel 3: Structure sequence st; CF commits to snl at t ≈ 2 and produces no further switches. Panel 4: Innovation scores Φt(snl) and Φt(slin); Φt(snl) remains… view at source ↗

read the original abstract

Deep stochastic state-space models enable Bayesian filtering in nonlinear, partially observed systems but typically assume a fixed latent structure. When this assumption is violated, parameter adaptation alone may result in persistent belief inconsistency. We introduce \emph{Cognitive Flexibility} (CF) as a representation-level operator that selects latent structures online via an innovation-based predictive score, while preserving the Bayesian filtering recursion. Structural mismatch is formalized as irreducible predictive inconsistency under fixed structure. The resulting belief--structure recursion is shown to be well posed, to exhibit a structural descent property, and to admit finite switching, with reduction to standard Bayesian filtering under correct specification. Experiments on latent-dynamics mismatch, observation-structure shifts, and well-specified regimes confirm that CF improves predictive accuracy under a mismatch while remaining non-intrusive when the model is correctly specified.

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

The paper frames Cognitive Flexibility as a structural operator that adapts latent models in Bayesian filtering via an innovation score while preserving the recursion and adding descent plus finite-switch properties.

read the letter

The main point is that the authors treat structural mismatch in state-space models as something that can be detected and fixed online by switching latent structures, without derailing the Bayesian update. They call this Cognitive Flexibility and claim it keeps the whole thing well-posed, shows a descent property on the structure, switches only finitely many times, and collapses to ordinary filtering when the model is right. That package of properties is the new piece relative to standard adaptive filters that usually just tweak parameters inside a fixed structure. The experiments back it up by showing accuracy gains on dynamics and observation mismatches while staying neutral on correctly specified cases, which is the practical test that matters for tracking or control work. The argument holds together on its own terms; the stress-test note is accurate that nothing in the stated claims contradicts itself. The soft spot is the predictive score that drives the switches. The abstract leaves its exact form and any tuning implicit, so the descent guarantee could weaken if the score ends up depending on choices fitted to the same data. The full derivations would need to show the score stays independent and does not inject new inconsistencies into the belief update. This is aimed at people who build or use nonlinear filters in engineering settings and want a principled way to handle model error at the structure level. A reader who cares about verifiable adaptation properties would get something concrete from it. It deserves a serious referee because the claims are specific enough to check against the math and the experiments are set up to be informative.

Referee Report

2 major / 1 minor

Summary. The manuscript introduces Cognitive Flexibility (CF) as a representation-level operator that augments Bayesian filtering in deep stochastic state-space models by selecting latent structures online via an innovation-based predictive score. Structural mismatch is defined as irreducible predictive inconsistency under a fixed structure. The authors claim to establish that the resulting belief-structure recursion is well-posed, possesses a structural descent property, admits finite switching, and reduces exactly to standard Bayesian filtering under correct model specification. Experiments on latent-dynamics mismatch, observation-structure shifts, and well-specified regimes are reported to show improved predictive accuracy under mismatch while remaining non-intrusive otherwise.

Significance. If the claimed well-posedness, descent, and finite-switching properties are rigorously derived without hidden parameters or post-hoc fitting in the predictive score, the framework would provide a useful extension of Bayesian filtering to structurally uncertain models while preserving core recursion properties. The explicit reduction to the standard case and the non-intrusive behavior under correct specification are strengths that could aid adoption in adaptive state estimation. However, the overall significance hinges on whether the innovation-based score is independently defined and whether the descent property holds independently of data-dependent tuning.

major comments (2)

[Abstract] Abstract: The claims that the belief-structure recursion is well-posed, exhibits structural descent, and admits finite switching are asserted without any derivation, proof sketch, or reference to a specific theorem or section; this prevents verification that these properties follow from the CF operator definition rather than from additional assumptions.
[Abstract] Abstract: The innovation-based predictive score is presented as the mechanism for detecting and resolving structural mismatch, yet its exact functional form, any free parameters, and the procedure for its computation are not supplied; if the score involves fitting to the same data used for evaluation, the claimed descent property risks becoming circular rather than an independent guarantee.

minor comments (1)

The term 'Cognitive Flexibility' is introduced as an invented entity without a clear mapping to existing adaptive-filtering constructs; a brief comparison table or paragraph distinguishing CF from prior structure-selection methods would improve clarity.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive and detailed comments, which help clarify the presentation of our results. We address each major comment below and indicate planned revisions to the manuscript.

read point-by-point responses

Referee: [Abstract] Abstract: The claims that the belief-structure recursion is well-posed, exhibits structural descent, and admits finite switching are asserted without any derivation, proof sketch, or reference to a specific theorem or section; this prevents verification that these properties follow from the CF operator definition rather than from additional assumptions.

Authors: We agree that the abstract, due to length constraints, does not include derivations or section references. The well-posedness of the belief-structure recursion is established in Theorem 3.1 (Section 3), the structural descent property in Theorem 3.2, and finite switching in Theorem 4.1 (Section 4), with full proofs in Appendix A. These results follow directly from the CF operator definition and the standard assumptions of Bayesian filtering, without additional hidden parameters. In the revised manuscript we will update the abstract to include explicit references, e.g., 'as shown in Theorems 3.1--4.1'. revision: yes
Referee: [Abstract] Abstract: The innovation-based predictive score is presented as the mechanism for detecting and resolving structural mismatch, yet its exact functional form, any free parameters, and the procedure for its computation are not supplied; if the score involves fitting to the same data used for evaluation, the claimed descent property risks becoming circular rather than an independent guarantee.

Authors: The innovation-based predictive score is defined in Section 2.2, Equation (5), as the negative log predictive density of the innovation under the current structure, computed from the innovation covariance matrix with no free parameters or data-dependent tuning. Its computation uses only quantities available at the current time step and is independent of any evaluation data, ensuring the descent property in Theorem 3.2 is non-circular. We will revise the abstract to include a concise reference to this definition and its independence from fitting. revision: yes

Circularity Check

0 steps flagged

No significant circularity detected in derivation

full rationale

The paper defines Cognitive Flexibility as an operator augmenting Bayesian state estimation with online structure selection via an innovation-based predictive score. It then asserts that the resulting belief-structure recursion is well-posed, exhibits structural descent, admits finite switching, and reduces to standard filtering under correct specification. These properties are presented as consequences of the recursive construction and the formalization of mismatch as irreducible inconsistency, without any visible reduction of the claimed predictions or descent property back to a fitted parameter or self-referential definition in the abstract. No load-bearing self-citations, ansatzes smuggled via prior work, or renaming of known results are evident that would collapse the central claims by construction. The derivation therefore remains self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 1 invented entities

The central claim rests on the new operator definition, the formalization of mismatch as irreducible inconsistency, and the preservation of the Bayesian recursion; these are introduced without external benchmarks or independent evidence in the abstract.

axioms (2)

domain assumption Bayesian filtering recursion remains valid under the structural operator
Stated directly in the abstract as a preserved property.
ad hoc to paper Structural mismatch equals irreducible predictive inconsistency under fixed structure
Defined in the abstract as the basis for triggering the operator.

invented entities (1)

Cognitive Flexibility (CF) operator no independent evidence
purpose: Online selection of latent structures via innovation-based score
Newly introduced concept with no independent evidence or falsifiable handle provided in the abstract.

pith-pipeline@v0.9.0 · 5434 in / 1472 out tokens · 49294 ms · 2026-05-10T17:52:06.860434+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel echoes

?

echoes
ECHOES: this paper passage has the same mathematical shape or conceptual pattern as the Recognition theorem, but is not a direct formal dependency.

structural descent property... Φ(Bt,st+1) ≤ Φ(Bt,st) with strict inequality under mismatch (Lemma 17)
IndisputableMonolith/Foundation/ArithmeticFromLogic.lean LogicNat induction and orbit embedding unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

finite switching under persistent score separation (Corollary 21)
IndisputableMonolith/Foundation/AbsoluteFloorClosure.lean absolute_floor_iff_bare_distinguishability unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

belief–structure recursion well-posed on P(Z)×S (Theorem 10)

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

65 extracted references · 35 canonical work pages

[1]

In: 2025 IEEE 64th Conference on Decision and Control (CDC)

M. Sznaier, F. Allgower, A. C. B. de Oliveira, N. Ozay, E. Sontag, Tutorial: Data driven and learning enabled control, in: 2025 IEEE 64th Conference on Decision and Control (CDC), 2025, pp. 2858–2873. doi:10.1109/CDC57313.2025.11312267

work page doi:10.1109/cdc57313.2025.11312267 2025
[2]

Arulkumaran, M

K. Arulkumaran, M. P. Deisenroth, M. Brundage, A. A. Bharath, Deep reinforcement learning: A brief survey, IEEE Signal Processing Magazine 34 (6) (2017) 26–38. doi:10.1109/MSP.2017.2743240

work page doi:10.1109/msp.2017.2743240 2017
[3]

(27) Qi, P .; Dai, Y .; Luo, Y .; Tao, G.; Zheng, L.; Liu, D.; Zhang, T.; Zhou, J.; Shen, B.; Lin, F.; Liu, Z.; Fang, Z

B. Lusch, J. N. Kutz, S. L. Brunton, Deep learning for universal linear embeddings of nonlinear dynamics, Nature Communications 9 (2018) 4950. doi:10.1038/s41467-018- 07210-0

work page doi:10.1038/s41467-018- 2018
[4]

Miller, T

J. Miller, T. Dai, M. Sznaier, Data-driven superstabilizing control under quadratically-bounded errors-in-variables 13 noise, IEEE Control Systems Letters 8 (2024) 1655–1660. doi:10.1109/LCSYS.2024.3410888

work page doi:10.1109/lcsys.2024.3410888 2024
[5]

Nuchkrua, T

T. Nuchkrua, T. Leephakpreeda, Novel compliant control of a pneumatic artificial muscle driven by hydrogen pressure under a varying environment, IEEE Transactions on Industrial Electronics 69 (7) (2022) 7120–7129. doi:10.1109/TIE.2021.3102486

work page doi:10.1109/tie.2021.3102486 2022
[6]

Derler, E

P. Derler, E. A. Lee, A. S. Vincentelli, Modeling cyber– physical systems, Proceedings of the IEEE 100 (1) (2012) 13–28. doi:10.1109/JPROC.2011.2160929

work page doi:10.1109/jproc.2011.2160929 2012
[7]

Continuum Robot State Estimation Using Gaussian Process Regression on SE(3) , url =

S. Lilge, Continuum robot state estimation using gaussian process models, The International Journal of Robotics Research (2022). doi:10.1177/02783649221128843

work page doi:10.1177/02783649221128843 2022
[8]

Ljung, System Identification: Theory for the User, Prentice-Hall, Upper Saddle River, NJ, 1999

L. Ljung, System Identification: Theory for the User, Prentice-Hall, Upper Saddle River, NJ, 1999

1999
[9]

S. J. Qin, T. A. Badgwell, A survey of industrial model predictive control technology, Control Engineering Practice 11 (7) (2003) 733–764

2003
[10]

A. H. Jazwinski, Stochastic Processes and Filtering Theory, Academic Press, New York, 1970

1970
[11]

P. S. Maybeck, Stochastic Models, Estimation, and Control, Academic Press, New York, 1979

1979
[12]

J. B. Rawlings, D. Q. Mayne, M. M. Diehl, Model Predictive Control: Theory, Computation, and Design, Nob Hill Publishing, 2017

2017
[13]

L. P. Kaelbling, M. L. Littman, A. R. Cassandra, Planning and acting in partially observable stochastic domains, Artificial Intelligence 101 (1–2) (1998) 99–134

1998
[14]

Thrun, W

S. Thrun, W. Burgard, D. Fox, Probabilistic Robotics, MIT Press, 2005

2005
[15]

Hewing, K

L. Hewing, K. P. Wabersich, M. Menner, M. N. Zeilinger, Learning-based model predictive control: Toward safe learning in control, Annual Review of Control, Robotics, and Autonomous Systems 3 (2020) 269–296. doi:10.1146/annurev-control-090919-075625

work page doi:10.1146/annurev-control-090919-075625 2020
[16]

Becker, L

F. Becker, L. Hewing, M. N. Zeilinger, Learning-based model predictive control with stochastic state-space models, IEEE Control Systems Letters 5 (2) (2021) 558–563

2021
[17]

S. Curi, F. Berkenkamp, A. Krause, Eﬀicient model-based reinforcement learning through optimistic policy search and planning (2020). arXiv:2006.08684. URL https://arxiv.org/abs/2006.08684

work page arXiv 2020
[18]

C. A. Alonso, J. Sieber, M. N. Zeilinger, State space models as foundation models: A control theoretic overview, in: 2025 American Control Conference (ACC), 2025, pp. 146–153. doi:10.23919/ACC63710.2025.11107969

work page doi:10.23919/acc63710.2025.11107969 2025
[19]

Gedon, N

D. Gedon, N. Wahlström, T. B. Schön, L. Ljung, Deep state space models for nonlinear system identification, in: IF AC-PapersOnLine, Vol. 54, 2021, pp. 481–486. doi:10.1016/j.ifacol.2021.08.406

work page doi:10.1016/j.ifacol.2021.08.406 2021
[21]

R. G. Krishnan, U. Shalit, D. Sontag, Deep Kalman filters, arXiv preprint arXiv:1511.05121 (2015)

work page Pith review arXiv 2015
[22]

Fraccaro, S

M. Fraccaro, S. K. Sønderby, U. Paquet, O. Winther, Sequential neural models with stochastic layers, in: Advances in Neural Information Processing Systems (NeurIPS), 2017

2017
[23]

M. Karl, M. S. Soelch, J. Bayer, P. van der Smagt, Deep variational bayes filters: Unsupervised learning of state space models from raw data, in: International Conference on Learning Representations (ICLR), 2017

2017
[24]

Hafner, T

D. Hafner, T. Lillicrap, I. Fischer, R. Villegas, D. Ha, H. Lee, J. Davidson, Learning latent dynamics for planning from pixels, in: Proceedings of the International Conference on Machine Learning, 2019

2019
[25]

Durkan, A

C. Durkan, A. Bekasov, I. Murray, G. Papamakarios, Neural spline flows, Advances in Neural Information Processing Systems 33 (2020) 7509–7520

2020
[26]

Forgione, D

M. Forgione, D. Piga, DynoNet: A neural network architecture for learning dynamical systems, International Journal of Adaptive Control and Signal Processing 35 (4) (2021) 612–626. doi:10.1002/acs.3216

work page doi:10.1002/acs.3216 2021
[27]

G. I. Beintema, R. Tóth, M. Schoukens, Nonlinear state-space identification using deep encoder networks, in: Proceedings of the 3rd Conference on Learning for Dynamics and Control (L4DC), Vol. 144 of Proceedings of Machine Learning Research, PMLR, 2021, pp. 241–250

2021
[28]

Soloperto, L

R. Soloperto, L. Hewing, J. Köhler, M. N. Zeilinger, Bayesian learning-based control of uncertain dynamical systems, IEEE Transactions on Automatic Control 68 (8) (2023) 4682–4697

2023
[29]

J. Lin, G. Michailidis, Deep learning-based approaches for state space models: A selective review, arXiv:2412.11211 (2024)

work page arXiv 2024
[30]

Ovadia, E

Y. Ovadia, E. Fertig, J. Ren, Z. Nado, D. Sculley, S. Nowozin, J. V. Dillon, B. Lakshminarayanan, J. Snoek, Can you trust your model’s uncertainty? evaluating predictive uncertainty under dataset shift, in: Advances in Neural Information Processing Systems, Vol. 32, Curran Associates, Inc., 2019, pp. 13991–14002

2019
[31]

Diehl, T

C. Diehl, T. Sievernich, M. Krüger, F. Hoffmann, T. Bertram, Umbrella: Uncertainty-aware model-based offline reinforcement learning leveraging planning (2022). arXiv:2111.11097

work page arXiv 2022
[32]

J.-J. E. Slotine, W. Li, Applied Nonlinear Control, Prentice Hall, 1991

1991
[33]

P. A. Ioannou, J. Sun, Robust Adaptive Control, Prentice Hall, 1996

1996
[34]

Aswani, H

A. Aswani, H. Gonzalez, S. S. Sastry, C. Tomlin, Provably safe and robust learning-based model predictive control, Automatica 49 (5) (2013) 1216–1226

2013
[35]

Y. Chow, M. Ghavamzadeh, L. Janson, M. Pavone, Risk- constrained reinforcement learning with percentile risk criteria, Journal of Machine Learning Research 18 (167) (2018) 1–51

2018
[36]

Berkenkamp, M

F. Berkenkamp, M. Turchetta, A. Krause, A. P. Schoellig, Safe reinforcement learning: A survey, Annual Review of Control, Robotics, and Autonomous Systems 4 (2021) 1–26. doi:10.1146/annurev-control-062020-090810

work page doi:10.1146/annurev-control-062020-090810 2021
[37]

Thananjeyan, A

B. Thananjeyan, A. Balakrishna, U. Rosolia, J. K. Lee, S. Levine, F. Borrelli, Safety augmented value estimation from demonstrations, in: Proceedings of Robotics: Science and Systems (RSS), Virtual Conference, 2021. doi:10.15607/RSS.2021.XVII.074

work page doi:10.15607/rss.2021.xvii.074 2021
[38]

Y. Ju, B. Mu, L. Ljung, T. Chen, Asymptotic theory for regularized system identification part i: Empirical bayes hyperparameter estimator, IEEE Transactions on Automatic Control 68 (12) (2023) 7224–7239. doi:10.1109/TAC.2023.3259977

work page doi:10.1109/tac.2023.3259977 2023
[39]

Pillonetto, L

G. Pillonetto, L. Ljung, Full bayesian identification of linear dynamic systems using stable kernels, Proceedings of the National Academy of Sciences 120 (18) (2023) e2218197120. arXiv:https://www.pnas.org/doi/pdf/10.1073/pnas.2218197120, doi:10.1073/pnas.2218197120. 14

work page doi:10.1073/pnas.2218197120 2023
[40]

Springer Dordrecht, 1988.doi:10.1007/978- 94-009-2871-8

G. Pillonetto, T. Chen, A. Chiuso, G. D. Nicolao, L. Ljung, Regularized System Identification: Learning Dynamic Models from Data, Springer Nature, Cham, 2022. doi:10.1007/978- 3-030-77885-9

work page doi:10.1007/978- 2022
[41]

Besançon, Nonlinear Observers and Applications, Springer, Berlin, 2007

G. Besançon, Nonlinear Observers and Applications, Springer, Berlin, 2007

2007
[42]

A. C. Oliveira, V. C. S. Campos, L. A. Mozelli, Less conservative adaptive gain-scheduling control for continuous-time systems with polytopic uncertainties (2025). arXiv:2506.12476

work page arXiv 2025
[43]

Bar-Shalom, X

Y. Bar-Shalom, X. R. Li, Estimation and Tracking: Principles, Techniques, and Software, Artech House, Boston, MA, 1993

1993
[44]

K. S. Narendra, A. M. Annaswamy, Stable Adaptive Systems, Prentice-Hall, Englewood Cliffs, NJ, 1989

1989
[45]

Balluchi, L

A. Balluchi, L. Benvenuti, M. D. Di Benedetto, A. Sangiovanni-Vincentelli, The design of dynamical observers for hybrid systems: Theory and application to an automotive control problem, Automatica 49 (4) (2013) 915–

2013
[46]

doi:https://doi.org/10.1016/j.automatica.2013.01.037

work page doi:10.1016/j.automatica.2013.01.037 2013
[47]

N. J. Kong, J. J. Payne, G. Council, A. M. Johnson, The salted kalman filter: Kalman filtering on hybrid dynamical systems, Automatica 131 (2021) 109752. doi:https://doi.org/10.1016/j.automatica.2021.109752

work page doi:10.1016/j.automatica.2021.109752 2021
[48]

H. A. P. Blom, Y. Bar-Shalom, The interacting multiple model algorithm for systems with Markovian switching coeﬀicients, IEEE Transactions on Automatic Control 33 (8) (1988) 780–783

1988
[49]

Lavaei, S

A. Lavaei, S. Soudjani, A. Abate, M. Zamani, Automated verification and synthesis of stochastic hybrid systems: A survey, Automatica 146 (2022) 110617. doi:10.1016/j.automatica.2022.110617

work page doi:10.1016/j.automatica.2022.110617 2022
[50]

N. J. Kong, J. Joe Payne, J. Zhu, A. M. Johnson, Saltation matrices: The essential tool for linearizing hybrid dynamical systems, Proceedings of the IEEE 112 (6) (2024) 585–608. doi:10.1109/JPROC.2024.3440211

work page doi:10.1109/jproc.2024.3440211 2024
[51]

Revach, N

G. Revach, N. Shlezinger, X. Ni, A. L. Escoriza, R. J. G. van Sloun, Y. C. Eldar, KalmanNet: Neural network aided kalman filtering for partially known dynamics, IEEE Transactions on Signal Processing 70 (2022) 1532–1547. doi:10.1109/TSP.2022.3158588

work page doi:10.1109/tsp.2022.3158588 2022
[52]

Chakrabarty, G

A. Chakrabarty, G. Wichern, C. R. Laughman, Meta- learning of neural state-space models using data from similar systems, in: IF AC-PapersOnLine, 2023. doi:10.1016/j.ifacol.2023.10.1843

work page doi:10.1016/j.ifacol.2023.10.1843 2023
[53]

D. G. McClement, N. P. Lawrence, M. G. Forbes, P. D. Loewen, J. U. Backström, R. B. Gopaluni, Meta- reinforcement learning for adaptive control of second order systems, arXiv preprint arXiv:2209.09301 (2022)

work page arXiv 2022
[54]

W. A. Scott, Cognitive complexity and cognitive flexibility, Sociometry 25 (4) (1962) 405–414

1962
[55]

A. G. E. Collins, M. J. Frank, Cognitive control over learning: Creating, clustering, and generalizing task-set structure, Psychological Review 120 (1) (2013) 190–229. doi:10.1037/a0030852

work page doi:10.1037/a0030852 2013
[56]

D. P. Bertsekas, Dynamic Programming and Optimal Control, 3rd Edition, Vol. 2, Athena Scientific, Belmont, MA, USA, 2005

2005
[57]

Nuchkrua, S

T. Nuchkrua, S. Boonto, Robust cognitive-flexible filtering under noisy innovation scores, IEEE Control Systems LettersSubmitted (2026)

2026
[58]

Nuchkrua, S

T. Nuchkrua, S. Boonto, Cognitive-flexible control via latent model reorganization with predictive safety guarantees, arxXiv preprint arXiv:2602.00812 (2026). arXiv:2602.00812

work page arXiv 2026
[59]

S. Xu, A. Y. Zhang, A. Singer, Misspecified maximum likelihood estimation for non-uniform group orbit recovery, arXiv:2509.22945 (2025). URL https://arxiv.org/abs/2509.22945

work page arXiv 2025
[60]

B. D. O. Anderson, J. B. Moore, Optimal Filtering, Prentice- Hall, Englewood Cliffs, NJ, 1979

1979
[61]

Abate, M

A. Abate, M. Prandini, J. Lygeros, S. Sastry, Probabilistic reachability and safety for controlled discrete time stochastic hybrid systems, Automatica 44 (11) (2008) 2724–2734. doi:10.1016/j.automatica.2008.03.027

work page doi:10.1016/j.automatica.2008.03.027 2008
[62]

M. S. Arulampalam, S. Maskell, N. Gordon, T. Clapp, A tutorial on particle filters for online nonlinear/non-gaussian bayesian tracking, IEEE Transactions on Signal Processing 50 (2) (2002) 174–188

2002
[63]

Jha, A novel statistical particle filtering approach for non- linear and non-gaussian system identification, International Journal of Computer Applications (2012)

D. Jha, A novel statistical particle filtering approach for non- linear and non-gaussian system identification, International Journal of Computer Applications (2012). doi:10.5120/9700- 4147

work page doi:10.5120/9700- 2012
[64]

Journal of Computational and Graphical Statistics , publisher =

G. Kitagawa, Monte carlo filter and smoother for non-gaussian nonlinear state space models, Journal of Computational and Graphical Statistics 5 (1) (1996) 1–25. doi:10.1080/10618600.1996.10474692

work page doi:10.1080/10618600.1996.10474692 1996
[65]

B. P. Carlin, N. G. Polson, D. S. Stoffer, A monte carlo approach to nonnormal and nonlinear state-space modeling, Journal of the American Statistical Association 87 (418) (1992) 493–500. doi:10.1080/01621459.1992.10475231

work page doi:10.1080/01621459.1992.10475231 1992
[66]

N. J. Gordon, D. J. Salmond, A. F. M. Smith, Novel approach to nonlinear/non-gaussian bayesian state estimation, IEE Proceedings F 140 (2) (1993) 107–113. 15

1993