pith. sign in

arxiv: 2607.02050 · v1 · pith:3QAUMQFWnew · submitted 2026-07-02 · 💻 cs.LG · cs.SY· eess.SY

A Memory Efficient Unified Algorithm for Online Learning of Linear Dynamical Systems

Pith reviewed 2026-07-03 17:09 UTC · model grok-4.3

classification 💻 cs.LG cs.SYeess.SY
keywords online learninglinear dynamical systemsregret minimizationinstability complexitymemory efficient algorithmspredictionstabilizationfilter based predictors
0
0 comments X

The pith

A unified online algorithm predicts any linear dynamical system using only O~(k) parameters when instability complexity is low.

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

The paper shows that for linear dynamical systems where only a small number k of modes are unstable, it is possible to perform online prediction with sublinear regret using an algorithm whose memory and parameters scale with k instead of the full hidden state dimension. This matters because the ability to stabilize an unknown system from observations hinges on having good predictions, and the authors argue that stabilization is only feasible without exponential control effort when k is small. Their unified algorithm covers all cases in this regime, such as non-diagonalizable systems or those with complex eigenvalues. They support this with a lower bound proving that k is the right complexity measure.

Core claim

Within the regime of systems with low instability complexity k, we introduce a unified online algorithm that handles every LDS including non-diagonalizable systems with complex or exploding modes with a learnable parameter count of O~(k). We prove sublinear regret and that any filter-based predictor needs at least k filters. This regime is where stabilization from observations is plausible.

What carries the argument

Unified online algorithm adapting parameter count to instability complexity k

Load-bearing premise

The premise that systems with high instability complexity cannot be stabilized without exponentially large controls.

What would settle it

Finding a high-dimensional LDS with small k on which the algorithm fails to achieve sublinear regret with O~(k) parameters, or a system with large k that can be stabilized with bounded controls.

read the original abstract

Motivated by the challenge of stabilizing a general unknown linear dynamical system (LDS) from observations, we study the natural prerequisite of online prediction. Our goal is to achieve sublinear regret with a memory footprint that adapts to the intrinsic complexity of the dynamics rather than the full hidden -- state dimension. We focus on the practically central regime of systems with low instability complexity -- eigenvalues outside the real stable interval that do not decay rapidly, together with non-semisimple modes-potentially embedded in an otherwise stable real spectrum of much higher dimension; we write $k$ for this count. This regime is the primary setting in which stabilization is plausible: we show that many systems with high instability complexity cannot be stabilized without exponentially large controls. Thus, prediction is meaningful for stabilization precisely when the instability complexity is small. Within this regime, we introduce a unified online algorithm that handles every LDS (including non-diagonalizable systems with complex or exploding modes) with a learnable parameter count of $\widetilde{O}(k)$. Finally, we prove a lower bound showing that $k$ is a valid complexity measure: any filter-based predictor needs at least $k$ filters. Experiments corroborate our theory: on a high-dimensional system, our predictor sharply outperforms prior methods at an equal parameter budget.

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.

Referee Report

0 major / 3 minor

Summary. The paper studies online prediction as a prerequisite for stabilizing unknown linear dynamical systems (LDS). It focuses on the regime of low instability complexity k (eigenvalues outside the stable interval that do not decay rapidly, plus non-semisimple modes). Within this regime it claims a single online algorithm that achieves sublinear regret for every LDS—including non-diagonalizable systems with complex or exploding modes—while using only ilde{O}(k) learnable parameters. It also proves a matching lower bound that any filter-based predictor requires at least k filters, shows that high-k systems cannot be stabilized without exponentially large controls, and reports experiments in which the new predictor outperforms prior methods on a high-dimensional system at equal parameter budget.

Significance. If the central claims hold, the work would be significant for online learning and adaptive control: it supplies a memory-efficient, unified method whose complexity tracks intrinsic instability rather than ambient state dimension, together with a matching lower bound that validates k as a complexity measure. The explicit handling of non-diagonalizable, complex, and exploding modes and the link between prediction and stabilization plausibility are notable strengths.

minor comments (3)
  1. The abstract states that experiments corroborate the theory on a high-dimensional system, but supplies no details on system dimension, choice of k, baselines, or performance metrics; these should be summarized with a table or figure reference in the main text.
  2. The notation ilde{O}(k) for the learnable parameter count is used without an explicit definition of the hidden constants or the precise dependence on other problem parameters (e.g., time horizon, noise variance).
  3. The lower-bound statement that “any filter-based predictor needs at least k filters” should clarify whether the result applies only to a specific class of filters or to all linear filters; a brief proof sketch or reference to the relevant theorem would help.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for the positive review, accurate summary of our contributions, and recommendation of minor revision. No major comments were raised in the report.

Circularity Check

0 steps flagged

No significant circularity identified

full rationale

The derivation chain in the abstract presents an algorithm achieving sublinear regret with ilde{O}(k) parameters for low-instability-complexity LDS, together with an independent lower bound that any filter-based predictor requires at least k filters. No self-definitional reduction, fitted-input prediction, or load-bearing self-citation is visible; the regime definition, algorithm, and complexity lower bound are stated as separate results without equations or citations that collapse one into the other by construction. The paper is self-contained against external benchmarks on this point.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Review based solely on abstract; no explicit free parameters, invented entities, or additional axioms beyond the stated domain assumption about high-k stabilizability are visible.

axioms (1)
  • domain assumption Many systems with high instability complexity cannot be stabilized without exponentially large controls.
    Invoked to justify focusing on the low-k regime as the setting where prediction is meaningful for stabilization.

pith-pipeline@v0.9.1-grok · 5765 in / 1149 out tokens · 24875 ms · 2026-07-03T17:09:59.378939+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

39 extracted references · 39 canonical work pages

  1. [1]

    2026 , eprint=

    Introduction to Online Control , author=. 2026 , eprint=

  2. [2]

    White , title =

    John S. White , title =. The Annals of Mathematical Statistics , number =. 1958 , doi =

  3. [3]

    Philosophical Transactions of the Royal Society of London , number =

    Cayley, Arthur , title =. Philosophical Transactions of the Royal Society of London , number =. 1858 , month =. doi:10.1098/rstl.1858.0002 , url =

  4. [4]

    2019 , eprint=

    Gradient Descent Learns Linear Dynamical Systems , author=. 2019 , eprint=

  5. [5]

    2017 , eprint=

    Learning Linear Dynamical Systems via Spectral Filtering , author=. 2017 , eprint=

  6. [6]

    2018 , eprint=

    Learning Without Mixing: Towards A Sharp Analysis of Linear System Identification , author=. 2018 , eprint=

  7. [7]

    2018 , eprint=

    Finite Time Identification in Unstable Linear Systems , author=. 2018 , eprint=

  8. [8]

    and Francis, Bruce , year =

    Doyle, John and Glover, Keith and Khargonekar, P. and Francis, Bruce , year =. State Space Solution to Standard. Automatic Control, IEEE Transactions on , doi =

  9. [9]

    at - Automatisierungstechnik , doi =

    Editorial: Effective construction of linear state-variable models from input/output functions , author =. at - Automatisierungstechnik , doi =. 1966 , lastchecked =

  10. [10]

    Kailath, Thomas , year =

  11. [11]

    2024 , eprint=

    A Tutorial on the Non-Asymptotic Theory of System Identification , author=. 2024 , eprint=

  12. [12]

    2020 , eprint=

    Online Learning of the Kalman Filter with Logarithmic Regret , author=. 2020 , eprint=

  13. [13]

    2020 , eprint=

    No-Regret Prediction in Marginally Stable Systems , author=. 2020 , eprint=

  14. [14]

    2022 , eprint=

    Learning to Control Linear Systems can be Hard , author=. 2022 , eprint=

  15. [15]

    Harry , year=

    Ng, Jerry and Asada, H. Harry , year=. Learned Lifted Linearization Applied to Unstable Dynamic Systems Enabled by Koopman Direct Encoding , volume=. doi:10.1109/lcsys.2022.3231641 , journal=

  16. [16]

    2018 , eprint=

    Theory and Algorithms for Forecasting Time Series , author=. 2018 , eprint=

  17. [17]

    2019 , eprint=

    Online Control with Adversarial Disturbances , author=. 2019 , eprint=

  18. [18]

    2021 , eprint=

    Black-Box Control for Linear Dynamical Systems , author=. 2021 , eprint=

  19. [19]

    2022 , eprint=

    Adaptive Regret for Control of Time-Varying Dynamics , author=. 2022 , eprint=

  20. [20]

    2020 , eprint=

    Finite-time Identification of Stable Linear Systems: Optimality of the Least-Squares Estimator , author=. 2020 , eprint=

  21. [21]

    2020 , eprint=

    Active Learning for Identification of Linear Dynamical Systems , author=. 2020 , eprint=

  22. [22]

    2023 , eprint=

    Learning the Kalman Filter with Fine-Grained Sample Complexity , author=. 2023 , eprint=

  23. [23]

    2018 , eprint=

    On the Sample Complexity of the Linear Quadratic Regulator , author=. 2018 , eprint=

  24. [24]

    2019 , eprint=

    Certainty Equivalence is Efficient for Linear Quadratic Control , author=. 2019 , eprint=

  25. [25]

    2020 , eprint=

    Logarithmic Regret for Adversarial Online Control , author=. 2020 , eprint=

  26. [26]

    2023 , eprint=

    Naive Exploration is Optimal for Online LQR , author=. 2023 , eprint=

  27. [27]

    Proceedings of the 24th Annual Conference on Learning Theory , pages =

    Regret Bounds for the Adaptive Control of Linear Quadratic Systems , author =. Proceedings of the 24th Annual Conference on Learning Theory , pages =. 2011 , editor =

  28. [28]

    2019 , eprint=

    Near optimal finite time identification of arbitrary linear dynamical systems , author=. 2019 , eprint=

  29. [29]

    2025 , eprint=

    Universal Learning of Nonlinear Dynamics , author=. 2025 , eprint=

  30. [30]

    Nathan and Brunton, Steven L

    Lusch, Bethany and Kutz, J. Nathan and Brunton, Steven L. , year=. Deep learning for universal linear embeddings of nonlinear dynamics , volume=. Nature Communications , publisher=. doi:10.1038/s41467-018-07210-0 , number=

  31. [31]

    Nathan and Brunton, Steven L

    Champion, Kathleen and Lusch, Bethany and Kutz, J. Nathan and Brunton, Steven L. , year=. Data-driven discovery of coordinates and governing equations , volume=. Proceedings of the National Academy of Sciences , publisher=. doi:10.1073/pnas.1906995116 , number=

  32. [32]

    2020 , eprint=

    Forecasting Sequential Data using Consistent Koopman Autoencoders , author=. 2020 , eprint=

  33. [33]

    Advances in Neural Information Processing Systems , volume=

    Neural Ordinary Differential Equations , author=. Advances in Neural Information Processing Systems , volume=

  34. [34]

    2019 , eprint=

    Neural Ordinary Differential Equations , author=. 2019 , eprint=

  35. [35]

    International Statistical Review , volume =

    Vovk, Volodya , title =. International Statistical Review , volume =. doi:https://doi.org/10.1111/j.1751-5823.2001.tb00457.x , url =. https://onlinelibrary.wiley.com/doi/pdf/10.1111/j.1751-5823.2001.tb00457.x , abstract =

  36. [36]

    2026 , eprint=

    Spectral Filtering for Complex Linear Dynamical Systems , author=. 2026 , eprint=

  37. [37]

    1996 , publisher=

    Subspace Identification for Linear Systems , author=. 1996 , publisher=

  38. [38]

    System Identification: Theory for the User, 2nd Edition (Ljung, L.; 1999) [On the Shelf] , volume =

    Simpkins, Charles , year =. System Identification: Theory for the User, 2nd Edition (Ljung, L.; 1999) [On the Shelf] , volume =. Robotics & Automation Magazine, IEEE , doi =

  39. [39]

    2024 , eprint=

    Provable Benefits of Complex Parameterizations for Structured State Space Models , author=. 2024 , eprint=