Home / Curated / Model Families / Hammerstein–Wiener model

Hammerstein–Wiener model

Model Families AIMech:HammersteinWienerModel DRAFT

A real-valued, single-input single-output, discrete-time deterministic dynamical model family formed by a memoryless static input function, a causal linear time-invariant block, and a memoryless static output function, with nonlinear behavior supplied by the static functions.

Identity Notes & Concept Scope Compared all 442 pinned AIO IDs, labels and scoped synonyms and 536 native IDs, names and aliases against both cascade names, punctuation variants, expanded descriptions and IDs: no normalized identity match. Ignored- and hidden-inclusive searches of the accessible recovery and previous worktrees returned 746 and zero lines. All 22 distinct matched texts were inspected; they concern existing two-block records and their review/report paths, or the pending combined-cascade cohort. Original Documents files were inaccessible during the initial search. After access returned, a full ignored- and hidden-inclusive search there at 2026-10-12T14:36:27Z found no candidate-name, expanded-description or reference hits; all 11 user files matched their preserved hashes. Read all 41 semantic AIO rows, 65 native name/definition neighbors, nine complete annotated CLASS nodes with incident edges, nine generated records and eight complete native neighbors. The existing Wiener/Hammerstein families have two blocks, their correlation records specify estimators, and NARX/NARMAX specify observed-history recurrences. A linear state-space layer is a neural component; GPR conditions a process prior. None supplies the proposed three-block identity. AIO GLM specifies a statistical link/variance structure, ARDL lag structure and ETS smoothing/trend/seasonality; neural layer/network classes require different scope. Shared linear special cases do not make the families equivalent. This is bounded to the inspected pin and accessible records.

Claims & Grounded Evidence

4 Claims

Mechanism #static-linear-static

Write the deterministic cascade as y=h(G(psi(u))), using h for the forward output function. The input function acts on each input sample, the linear block carries the dynamics, and h acts on its output. Equivalently, a finite-state linear block has x_next=A*x+B*psi(u) and y=h(C*x+D*psi(u)). With psi(v)=v², causal filter G=(1,2), h(v)=v+v³, inputs (1,2,-1) and zero prehistory, outputs are (2,222,738).

supports Source: Data-Driven Prediction and Control of Hammerstein-Wiener Systems with Implicit Gaussian Processes · Locator: Section 2, Figure 1 and Equation 1, PDF/printed p. 3, setting noise to zero and writing h=phi^(-1); curator finite example in research/discovery/three-block-cascade-boundaries.py

Scope #two-block-special-cases

An identity input function psi gives a Wiener model, and an identity output function h gives a Hammerstein model. If G is the identity, the model reduces to the static composition h(psi(u)); if both static functions are identities, it reduces to G. These inclusions do not make the ordered three-block family identical to either two-block family or to an identification algorithm.

supports Source: Data-Driven Prediction and Control of Hammerstein-Wiener Systems with Implicit Gaussian Processes · Locator: Introduction, Section 2 Equation 1 and Section 3.2 Remark 4, PDF/printed pp. 1,3,6; curator algebraic specializations in research/discovery/three-block-cascade-boundaries.py

Limitation #internal-gain-and-printed-formula

For nonzero a,b in the zero-state deterministic formulation, psi_new=a*psi, G_new=(b/a)*G and h_new(v)=h(v/b) preserve the response. With a=2,b=3 the finite example still gives (2,222,738). By contrast, the source’s printed Remark 3 transformation (a*psi,G/(a*b),b*h) gives (10/9,30,90) for the same example and does not generally preserve a nonlinear output function. The counterexample and corrected coordinate rescaling are curator derivations, not an author-issued correction.

supports Source: Data-Driven Prediction and Control of Hammerstein-Wiener Systems with Implicit Gaussian Processes · Locator: Section 2 Equation 1, PDF/printed p. 3; Section 3.2 Remark 3, p. 6; exact source-formula and coordinate-rescaling checks in research/discovery/three-block-cascade-boundaries.py

Scope #noise-location-and-predictor-scope

The inspected GP predictor places noise between G and the output function, giving h(z+e), and assumes a stable G and a monotone invertible output map. These are conditions of that particular noisy predictor, not requirements for evaluating every forward deterministic cascade. Noise before h is generally not independent additive noise after h: for h(z)=z+z³ and independent standard Gaussian e, E[h(z+e)]-h(z)=3z. The predictor’s virtual derivative observations promote monotonicity but do not guarantee it globally.

supports Source: Data-Driven Prediction and Control of Hammerstein-Wiener Systems with Implicit Gaussian Processes · Locator: Section 2 Figure 1, Equation 1 and Assumption 1, PDF/printed p. 3; Section 3.3, pp. 6–7; curator exact noise moments in research/discovery/three-block-cascade-boundaries.py

⚠️ Curated Limitations & Negative Results

This maintained record selects real-valued scalar discrete-time deterministic cascades. Noise models, multivariable, parallel, feedback, time-varying and neural variants require explicit additional scope. Finite examples use causal FIR blocks and zero prehistory; general linear blocks may have infinite impulse-response memory or specified nonzero initial states.
A memoryless function is a single-valued map on its stated domain. Forward composition does not in itself require inverse functions, monotonicity, Gaussian inputs, polynomial parameterization or a GP prior. Any use of inverse-coordinate predictors must state the valid domain and stronger assumptions.
Representation, minimality, internal identifiability, stable simulation and estimation accuracy are separate questions. Internal gain normalizations alone do not settle all ambiguities; overlapping special cases do not establish family equivalence.
The 2026 source is a preprint, and this record does not adopt its predictor as a general performance or control guarantee. The printed gain formula is contradicted by an explicit curator example. Multivariable Jacobian/determinant claims, GP uncertainty calibration and closed-loop guarantees remain separate audits.
All pages of the cited version were visually read. No model was trained, no empirical result was replicated and no paid provider was run. This record remains DRAFT; author self-review is not independent scientific approval.

Artificial Intelligence Ontology (AIO) Grounding

Mapped AIO Classes

AIO Target Class	Mapping Relation	Mapping Rationale
`AIO:TimeSeriesAnalysis` Time Series Analysis	related_match	This ordered dynamical model family represents temporal input-output relationships for system identification and analysis. The inspected AIO target is an analysis task, so the link is a model/task relationship rather than exact identity or an asserted subclass.

Evidence Sources & Primary Literature

1 Source

Source ID	Publication / Document	Type / Access	Reference & Checksum
`yin-mueller2026`	Data-Driven Prediction and Control of Hammerstein-Wiener Systems with Implicit Gaussian Processes Mingzhou Yin, Matthias A. Müller	primary paper full_text	ID: https://arxiv.org/abs/2501.15849v2 sha256:0d0701623f76…

Curation Provenance

Curator: Codex

Method: ai_assisted

Model: GPT-6

Created: 2026-10-12T14:34:20Z

Updated: 2026-10-12T14:42:40Z

Source File: data/curated/HammersteinWienerModel.yaml