{"id":"be790f36-37f5-461d-aaaa-ec1e47a266a5","arxiv_id":"2606.17509","paper_version":1,"verdict":"UNVERDICTED","confidence":"LOW","novelty_score":6.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":0,"one_line_summary":"A data-driven approach builds local exponentially input-to-state stabilizing controllers from noisy data per subsystem and composes them via small-gain conditions to achieve uniform global exponential stability for infinite LTI networks.","lead":"The paper describes a data-driven method to design stabilizing controllers for networks made of infinitely many unknown linear subsystems using noisy trajectory data from each one. A generalist might care because it addresses how to guarantee stability in very large interconnected systems like power grids or sensor arrays without needing exact mathematical models of every part.","discovery_kind":"new_method","skeptic_critique":{"model":"grok-4.3","headline":"No significant objection identified","rationale":"The reader's weakest_assumption matches the explicit conditional phrasing in the abstract. With the claim already caveated on the LMIs and small-gain condition, and no contradictory derivation visible, the argument does not contain an additional load-bearing gap beyond what the reader noted.","tokens_in":1651,"tokens_out":221,"duration_ms":22173,"concrete_test":"Re-derive the local LMI feasibility condition from the data matrices in the physical case study section; confirm that the reported noise bound still yields a feasible solution and that the resulting local gains satisfy the infinite-network small-gain inequality with a uniform margin.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is explicitly conditional on LMIs being feasible from the noisy data and on the compositional small-gain condition holding in the infinite-dimensional setting. The abstract states these prerequisites clearly and does not assert unconditional stability. No internal inconsistency appears in the stated argument structure (local data-driven eISS design followed by small-gain composition).","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.3","summary":"The paper proposes a direct data-driven method for controller synthesis of infinite networks composed of unknown linear time-invariant subsystems. Using a single set of noise-corrupted input-state trajectories collected from each subsystem, and provided that certain linear matrix inequalities hold, each subsystem is rendered exponentially input-to-state stable (eISS) by locally constructing an eISS control Lyapunov function together with an exponentially input-to-state stabilizing feedback controller. These local components are then composed under a compositional small-gain condition in infinite-dimensional spaces to obtain a global control Lyapunov function and an associated stabilizing controller, ensuring uniform global exponential stability of the infinite network. The approach is validated on a physical case study with unknown dynamics.","tokens_in":1695,"tokens_out":442,"duration_ms":17108,"significance":"If the LMI feasibility conditions from noisy data and the infinite-dimensional small-gain condition can be verified, the result provides a model-free route to stabilizing controllers for infinite networks, extending data-driven Lyapunov methods to compositional infinite-dimensional settings. The explicit conditioning on data-derived LMIs and the small-gain test is a strength, as is the use of a physical case study for validation.","major_comments":[],"minor_comments":[{"comment":"The case-study section should report the specific LMI feasibility outcomes, the computed local gains, and the numerical verification of the small-gain condition (including the value of the gain margin) so that readers can assess how close the design operates to the boundary of the assumptions.","section":"Case study"},{"comment":"Clarify the precise statement of the infinite-dimensional small-gain theorem invoked (reference and any modifications for the eISS setting) and confirm that the composition preserves the exponential decay rate uniformly across the network.","section":"Section on compositional small-gain condition"},{"comment":"The data-driven LMI formulation should explicitly state the noise bound assumption and how it enters the matrix inequality; if the bound is treated as a design parameter, note its effect on feasibility.","section":"Local controller design"}],"recommendation":"minor_revision","confidential_remarks":null},"author_rebuttal":{"model":"grok-4.3","summary":"We thank the referee for the positive summary, recognition of the significance of the data-driven compositional approach, and the recommendation of minor revision. No specific major comments were listed in the report.","responses":[],"tokens_in":1159,"tokens_out":56,"duration_ms":14008,"standing_objections":[]},"desk_editor":{"model":"grok-4.3","letter":"The central point is that you can stabilize an infinite network of unknown linear subsystems using only one set of noisy input-state trajectories per subsystem. Local LMIs yield eISS control Lyapunov functions and feedbacks, which are then put together via an infinite-dimensional small-gain condition for global uniform exponential stability.\n\nWhat is new here is the direct data-driven synthesis for the infinite case. Prior work handled finite networks, but this adapts the compositional small-gain argument to infinite dimensions while keeping the data-driven local step.\n\nIt does a good job stating the prerequisites explicitly, so the claims are conditional on the LMIs being solvable and the small-gain holding. That keeps it honest.\n\nThe soft spots are around the details we can't see yet. How the noise is bounded in the data-driven LMIs isn't shown in the abstract, and the case study validation isn't described, so it's unclear how tight the conditions are in practice or if there are hidden assumptions in the infinite setting. If those check out in the full text, the argument holds.\n\nThis paper is for control engineers and theorists focused on large-scale or networked systems. A reader working on data-driven methods for unknown dynamics would get value from the framework.\n\nIt deserves peer review to examine the proofs and the case study results.","headline":"This paper gives a data-driven way to stabilize infinite linear networks by designing local eISS controllers from noisy data and composing them with small-gain conditions.","tokens_in":2156,"tokens_out":336,"would_cite":false,"duration_ms":22679,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.3","headline":"A single set of noisy trajectories per subsystem yields local controllers that compose via small-gain to stabilize an entire infinite network.","keywords":["data-driven control","infinite networks","input-to-state stability","small-gain condition","linear matrix inequalities","control Lyapunov functions","exponential stability","unknown linear systems"],"falsifier":"Numerical or experimental observation that the closed-loop network fails to achieve uniform global exponential stability even though the local linear matrix inequalities are feasible and the small-gain condition is satisfied.","tokens_in":2570,"feed_emoji":"🔗","tokens_out":750,"duration_ms":22648,"temperature":0.7,"pith_summary":"The paper develops a direct data-driven approach to design stabilizing controllers for infinite networks of unknown linear time-invariant subsystems. From one collection of noise-corrupted input-state data per subsystem, linear matrix inequalities are solved to produce a local exponentially input-to-state stabilizing controller and associated control Lyapunov function for each unit. These local designs are then assembled under a small-gain condition formulated for infinite-dimensional spaces, which produces a global control Lyapunov function and feedback law that renders the whole network uniformly globally exponentially stable. A reader would care because many engineered systems consist of large numbers of similar interacting components whose exact models are unavailable, and the method shows how limited local measurements can still guarantee network-wide stability without centralized identification.","feed_headline":"Noisy local data composes into global stabilizer for infinite networks","feed_subtitle":"One trajectory set per subsystem produces local eISS controllers that satisfy an infinite-dimensional small-gain condition and guarantee uni","key_machinery":"The compositional small-gain condition in infinite-dimensional spaces that assembles local eISS control Lyapunov functions and controllers into a global stabilizing pair.","core_discovery":"Using a single set of noise-corrupted input-state trajectories collected from each subsystem, and provided that certain linear matrix inequalities hold, each subsystem is rendered exponentially input-to-state stable by locally constructing an eISS control Lyapunov function together with an exponentially input-to-state stabilizing feedback controller. These local components are composed under a compositional small-gain condition in infinite-dimensional spaces to obtain a global control Lyapunov function and an associated stabilizing controller, ensuring uniform global exponential stability of the infinite network.","pith_inferences":["The framework could be tested on finite truncations of increasing size to check whether stability margins remain uniform as the network grows.","If the small-gain parameters can be adjusted locally, the same data-driven procedure might extend to networks whose subsystems are only approximately identical.","One could examine whether the local linear matrix inequalities remain feasible under larger noise bounds, which would indicate practical robustness limits.","The approach points toward data-driven certification of stability in other distributed systems where exact interconnection strengths are also uncertain."],"forward_implications":["Each unknown linear subsystem admits an exponentially input-to-state stabilizing controller constructed solely from its own noisy data when the associated inequalities are solvable.","The infinite network reaches uniform global exponential stability once the local controllers satisfy the infinite-dimensional small-gain condition.","The resulting global controller requires only local state measurements and does not need a centralized model of the full network.","The same data set suffices both to certify local stability properties and to enable the global composition.","The method applies directly to physical systems whose dynamics are treated as unknown."],"fun_headline_variants":["Local noisy data stabilizes infinite networks","Single trajectories enable infinite network control","Data composes local eISS into global stability","Small-gain unifies data-based infinite stabilizers"],"cache_read_input_tokens":2112,"weakest_assumption_plain":"The linear matrix inequalities can be solved from the collected noisy trajectories to produce valid local controllers, and the small-gain condition holds for the infinite network.","fun_headline_variants_meta":{"raw":{"variants":["Local noisy data stabilizes infinite networks","Single trajectories enable infinite network control","Data composes local eISS into global stability","Small-gain unifies data-based infinite stabilizers"]},"model":"grok-4.3","cost_usd":0.005588,"raw_usage":{"total_tokens":2548,"prompt_tokens":572,"num_sources_used":0,"completion_tokens":51,"cost_in_usd_ticks":55878000,"prompt_tokens_details":{"text_tokens":572,"audio_tokens":0,"image_tokens":0,"cached_tokens":64},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":1925,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":572,"tokens_out":51,"duration_ms":19197,"temperature":1.0,"reasoning_tokens":1925,"cache_read_input_tokens":64,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-06-26T23:32:10.425642+00:00","model_set":{"reader":"grok-4.3"},"falsifier":"Numerical or experimental observation that the closed-loop network fails to achieve uniform global exponential stability even though the local linear matrix inequalities are feasible and the small-gain condition is satisfied.","supporting_citations":[],"review_version":1}