← All papers

Functorial Framework for Morphological Computing

Published: 2026-07-04

A

Functorial Framework for Morphological Computing: A Vectorial,

Topologically-Protected Solution to the Residue Number System

Reconstruction Paradox

Author: Rowan Brad Quni-Gudzinas

Contact: rowan.quni@outlook.com ORCID:

0009-0002-4317-5604 ISNI: 0000 0005 2645 6062

DOI: 10.5281/zenodo.17491258 **Publication

Date: 2025-10-31 Version:** 1.0

Abstract: This work introduces a formal framework

for morphological computing that resolves the longstanding

reconstruction paradox in Residue Number System (RNS) arithmetic by

leveraging the physics of topological quantum matter. Conventional

RNS-based accelerators are bottlenecked by the serial, computationally

expensive conversion of a parallel residue vector into a scalar integer

via the Chinese Remainder Theorem. We dissolve this bottleneck by

proposing a vectorial topological encoding scheme where the

computational result is the residue vector itself, physically embodied

as a vector of Chern numbers in a moiré superlattice. We address the

critical challenge of formal verification in physical computing by

constructing a structure-preserving functor from the algebraic category

of RNS computations to the physical category of adiabatic evolutions in

topological phases. This functor provides a formal guarantee of

computational correctness. The final result is read out directly via a

multi-terminal quantum Hall conductance measurement, which yields the

Chern vector without any algorithmic post-processing. This approach

establishes a new paradigm of computation-by-relaxation that is formally

verifiable, physically robust, and achieves end-to-end energy efficiency

by eliminating the reconstruction step entirely.

Keywords: morphological computing, Residue Number

Systems, topological quantum matter, functorial framework, category

theory, Chern insulator, moiré materials, quantum Hall effect, physical

computation, hardware-software co-design, energy-efficient computing

I. State of

the Art and Gap Identification

A. Summary of key prior work

The scholarly context for this work is situated at the confluence of

three distinct but complementary fields: morphological computing,

Residue Number Systems (RNS), and topological quantum matter. The

concept of morphological computing, where a system’s physical structure

contributes to computation, was advanced by Pfeifer & Bongard

(2007), but has been critically assessed as often lacking a rigorous

information-theoretic foundation that would distinguish it from trivial

physical dynamics (Müller & Hoffmann, 2017). This concern is

reinforced by the argument that without a formal framework to define the

relationship between a physical process and an abstract computation,

claims of physical computation risk becoming unverifiable (Horsman et

al., 2014). In parallel, RNS, as formalized by Szabo & Tanaka

(1967), offers an architecture for carry-free parallel arithmetic.

However, its practical application is persistently hindered by the

computational overhead of the Chinese Remainder Theorem (CRT)

reconstruction step, a bottleneck that remains a challenge even in

modern applications such as homomorphic encryption (Cheon et al., 2017).

Concurrently, the study of topological phases of matter has progressed

from theoretical proposals (Kitaev, 2003) to experimental realizations

of robust, topologically protected states in moiré materials (Sharpe et

al., 2019). While these systems provide stable, quantized invariants,

the literature lacks a clear protocol for encoding arbitrary

computational results into these invariants and reading them out

directly. Existing work on topological quantum computing has largely

focused on developing fault-tolerant quantum gates using non-Abelian

anyons, rather than addressing general-purpose integer arithmetic (Nayak

et al., 2008). The broader field of physical computation has long

explored analogues to Turing-completeness in continuous physical systems

but has not yet produced scalable, noise-resilient, and formally

verifiable architectures for integer arithmetic (MacLennan, 2004;

Siegelmann, 1999). Finally, the persistent demand for energy-efficient

computing in the post-Moore era highlights the limitations of existing

RNS accelerators, which have yet to overcome the systems-level overhead

imposed by the reconstruction step (Horowitz, 2014).

B. Identified open problems or tensions

The confluence of these fields reveals a set of interconnected and

unresolved problems. The primary issue is the reconstruction paradox,

where the inherent parallelism of RNS is nullified by the serial,

computationally intensive CRT reconstruction required to produce a

usable scalar output. This is directly linked to the verification

challenge: many proposals for physics-based computation remain

descriptive, lacking the formal mathematical structure, such as that

provided by category theory, needed to prove that a physical system

correctly implements a specified computation (Müller & Hoffmann,

2017; Horsman et al., 2014). Even if a computational state were encoded

in a topological invariant, a measurement problem persists, as no

established protocol allows for the direct measurement of such an

invariant to yield a final computational result without requiring

further algorithmic processing. This has led to a scalar encoding

failure, where attempts to map the CRT reconstruction onto a single,

weighted physical observable have proven unworkable due to non-physical

requirements and a misunderstanding of measurement principles. These

specific issues point to a deeper ontological mismatch: prior work

typically treats a physical system as a passive substrate on which an

abstract algorithm is executed, whereas this work proposes an inversion

where the computation is the natural dynamics of the physics itself.

Furthermore, existing topological computing frameworks exhibit a lack of

arithmetic universality, focusing on specialized quantum gate sets

rather than the general-purpose integer arithmetic needed for many

classical computing tasks. This disconnect is formalized by the absence

of categorical grounding, as no prior work has constructed a functorial

bridge to formally link an algebraic model of computation with a

category of topological quantum phases. Finally, these issues culminate

in a practical energy-latency tradeoff, where the benefits of RNS

arithmetic are offset by the costs of reconstruction, yielding no net

system-level advantage.

C. Positioned contribution of this work

This work presents a framework that directly addresses the identified

gaps by integrating these fields through a novel, formally grounded

approach. It resolves the reconstruction paradox by dissolving the

problem: a vectorial topological encoding is proposed where the

computational result is the vector of residues itself, rendered directly

as a physical observable, eliminating the need for a scalar

reconstruction. The verification challenge is met through a

categorically formalized framework that constructs an explicit,

structure-preserving functor from the category of RNS computations to

the category of topological physical evolutions, providing a formal

proof of correctness that responds to the critiques of Müller &

Hoffmann (2017). The measurement problem is solved by specifying an

experimentally feasible, direct multi-terminal measurement protocol

where the encoded Chern vector is read out as a set of quantized Hall

conductances, yielding the final result without post-processing. This

approach overcomes the scalar encoding failure by embracing a parallel

vector output that is native to both RNS and the physics of topological

matter. By grounding the framework in a computation-by-relaxation

paradigm, this work offers a concrete realization of morphological

computing that is non-trivial and formally verifiable, as called for by

Horsman et al. (2014). It establishes arithmetic universality for a

class of topological systems by demonstrating a mapping from any integer

arithmetic operation to a realizable physical protocol. Through an

ontological inversion, computation is redefined not as an abstract

process imposed on matter, but as the natural, deterministic relaxation

of a constrained physical system. By eliminating the reconstruction

bottleneck, this architecture achieves end-to-end energy efficiency,

breaking the longstanding energy-latency tradeoff that has limited the

utility of RNS accelerators.

II.

Theoretical Foundations of Morphological Computing

A. From symbolic to physical computation

This framework redefines computation not as the execution of a

logical sequence, but as the deterministic evolution of a constrained

physical system. The process begins with the preparation of a

high-energy initial state, which then naturally relaxes to a

minimum-energy ground state that physically encodes the computational

solution. This “computation-by-relaxation” paradigm leverages the

system’s natural physics to perform computational work, thereby avoiding

the von Neumann bottleneck and the energy costs associated with

transistor switching. The process can be understood as a physical

instantiation of coarse-graining, where microscopic degrees of freedom

self-organize into a stable, macroscopic state that represents the

computational result. This view is consistent with Landauer’s principle,

as the energy dissipated is fundamentally linked to the physical

relaxation process rather than the logical irreversibility of an

abstract gate model. The computational complexity of a problem is

encoded in the landscape of the system’s free energy functional, with

the solution path corresponding to a geodesic in the system’s state

space, a concept that aligns with the notion of thermodynamic depth

(Lloyd & Pagels, 1988). The input is encoded in the preparation of

the initial state, while the function to be computed is encoded in the

topology of the energy landscape, thus unifying program and data within

the physical substrate. The system functions as a dissipative structure,

where an external energy flow is used to prepare the system far from

equilibrium, and its subsequent relaxation toward equilibrium performs

the computation (Prigogine, 1967).

**B. Residue number systems as the optimal physical

logic**

Residue Number Systems provide the ideal mathematical structure for

this physical computing paradigm. The core advantage of RNS is its

inherent parallelism; arithmetic operations on each residue channel are

performed independently of all others, which eliminates the carry

propagation that fundamentally limits conventional binary arithmetic.

This mathematical independence maps directly and naturally onto a

physical architecture of decoupled subsystems, such as distinct regions

of a mesoscopic material, where each prime modulus corresponds to a

unique, topologically protected computational channel. The Chinese

Remainder Theorem guarantees a bijective mapping between a global

integer and its corresponding vector of local residues, ensuring that

the vector representation is both complete and unambiguous. The product

ring structure, \(\prod_{i=1}^k

\mathbb{Z}/p_i\mathbb{Z}\), is not merely a mathematical

convenience but is the formal expression of a physically decomposable

system, making RNS the natural logic for a modular hardware

implementation. The isomorphism of algebras guaranteed by the CRT

ensures that all essential ring-theoretic properties, such as

distributivity and associativity, are preserved in the vectorial

representation, providing a faithful encoding of the computation.

Furthermore, the use of small prime moduli provides a balance between

dynamic range, physical realizability, and intrinsic error detection, as

any error affecting a single channel produces an out-of-range vector

that is immediately detectable. RNS thus functions as a non-linear

error-detecting code, a property that is maintained and physically

grounded in the topological embodiment.

**C. Topological protection for robust information

encoding**

To ensure that the physical embodiment of RNS is robust, information

is encoded in global topological invariants that are intrinsically

resilient to local noise, defects, and thermal perturbations. The Chern

number, a global property of a system’s electronic band structure,

serves as the physical carrier of a residue value. Fractional Chern

Insulators (FCIs), which have been realized in moiré superlattices,

offer a suitable physical platform where the ground-state degeneracy and

associated Chern number can be engineered to be prime-bounded, thereby

creating a direct physical realization of an RNS residue channel. In

this architecture, the moiré pattern is not a passive substrate but an

active computational landscape whose electronic properties can be

programmed by tuning physical parameters like twist angle and external

electric fields. The topological degeneracy of the ground state provides

a natural Hilbert space for storing a residue, with each distinct Chern

sector functioning as a stable, noise-immune memory element. The

robustness of this encoding is physically guaranteed by the spectral gap

of the system, which exponentially suppresses fluctuations that could

cause unintended transitions between different topological sectors.

Information can be read out by leveraging the chiral edge states

characteristic of such systems, which carry a quantized current directly

proportional to the bulk Chern number, enabling a non-invasive,

contact-based measurement. This system exhibits topological quantum

order, where long-range quantum entanglement provides a deep, intrinsic

fault tolerance that protects the encoded information beyond the simple

energy barrier of the spectral gap (Wen, 2002).

**III. The Functorial Framework: A Formal Bridge from

Computation to Physics** |

**Appendix A: Formal Verification of the Functorial

Framework** |

The formal correctness of the proposed framework is established

through the following derivation. |

  1. Chinese Remainder Theorem: For a set of pairwise

coprime integers \(p1, \dots, pk\),

the map \(\phi: N \mapsto (N \bmod p_1, \dots,

N \bmod pk)\) defines a ring isomorphism \(\phi: \mathbb{Z}/M\mathbb{Z} \to \prod{i=1}^k

\mathbb{Z}/p_i\mathbb{Z}\), where \(M =

\prodi pi\). |

  1. Category RNS: The category RNS

is defined. Its objects are the rings \(\mathbb{Z}/p_i\mathbb{Z}\). Its morphisms

are ring homomorphisms \(f:

\mathbb{Z}/pi\mathbb{Z} \to \mathbb{Z}/pi\mathbb{Z}\) induced

by arithmetic operations (e.g., \(f(x) = x + a

\bmod p_i\)). |

  1. Category TopPh: The category

TopPh is defined. Its objects are pairs \((\mathcal{H}i, Hi)\), where \(\mathcal{H}_i\) is the ground-state Hilbert

space of an FCI with Chern numbers in \(\{0,

\dots, pi-1\}\), and \(Hi\) is

its Hamiltonian. Its morphisms are unitary operators \(U_f\) generated by adiabatic protocols,

given by: |

\[U_f = \mathcal{T}

\exp\left(-\frac{i}{\hbar} \int0^\tau Hi(t) \, dt\right) \quad

(A1)\] |

  1. Functor Construction (Objects): The functor

\(\mathcal{F}: \textbf{RNS} \to

\textbf{TopPh}\) is constructed on objects by the mapping \(\mathcal{F}(\mathbb{Z}/p_i\mathbb{Z}) =

(\mathcal{H}i, Hi)\). |

  1. Functor Construction (Morphisms): On morphisms,

for an arithmetic operation \(f(x) = x + a

\bmod pi\), the functor maps it to a unitary evolution \(\mathcal{F}(f) = Uf\), where \(U_f\) is generated by a Thouless pumping

protocol that shifts the Chern number by \(a\). |

  1. Identity Preservation: The functor preserves

identities. The identity morphism in RNS is \(f(x)=x\), which corresponds to a shift of

\(a=0\). The corresponding physical

protocol is a static Hamiltonian, which generates the identity unitary

\(\mathbb{I}{\mathcal{H}i}\). Thus,

\(\mathcal{F}(\mathrm{id}{\mathbb{Z}/pi\mathbb{Z}})

= \mathbb{I}{\mathcal{H}i}\). |

  1. Composition Preservation: The functor preserves

composition. For two morphisms \(f\)

and \(g\), the physical protocol for

\(g \circ f\) is the temporal

concatenation of the individual protocols. The resulting unitary

evolution is the product of the individual unitaries, \(Ug Uf\). Therefore, \(\mathcal{F}(g \circ f) = Ug Uf = \mathcal{F}(g)

\circ \mathcal{F}(f)\). |

  1. Conclusion: Since \(\mathcal{F}\) preserves both identities and

composition, it is a well-defined, structure-preserving functor. This

provides a formal guarantee that the physical system correctly

implements the algebra of RNS computations. |

  1. Universality: The framework is general. Any

other physical system that realizes the same algebraic structure (e.g.,

a photonic lattice) would be related to this moiré implementation by a

natural isomorphism, demonstrating the universality of the categorical

approach. |

  1. Extension to Multiplication: Ring homomorphisms

for multiplication (e.g., \(f_b(x) = b x \bmod

p_i\)) can be implemented via sequences of additions or other

non-linear adiabatic protocols, ensuring the full ring structure is

preserved under \(\mathcal{F}\). |

  1. Faithfulness: The functor is faithful. If \(\mathcal{F}(f) = \mathcal{F}(g)\), their

corresponding physical effects are identical. Since distinct arithmetic

operations produce distinct and measurable shifts in the Chern number,

this implies \(f = g\). |

  1. Fullness: The functor is full. Every adiabatic

protocol that shifts the Chern number by an integer \(a\) corresponds to the morphism \(fa(x) = x + a \bmod pi\) in

RNS. |

  1. Categorical Equivalence: Because the functor

\(\mathcal{F}\) is full, faithful, and

essentially surjective onto the relevant subcategory of

TopPh, it establishes an equivalence of categories

between RNS and its physical realization. |

Appendix

B: Verification of the Vectorial Encoding Scheme

The validity of the vectorial encoding and measurement scheme is

established as follows.

Bijective Mapping (CRT): From the Chinese

Remainder Theorem, the map \(\phi: N \mapsto

(r1, \dots, rk)\), where \(r_i = N

\bmod p_i\), is a bijection from the set of integers \(\{0, \dots, M-1\}\) to the product ring

\(\prod_{i=1}^k

\mathbb{Z}/p_i\mathbb{Z}\).

Physical State Representation: In the proposed

physical system, the computational state is characterized by the Chern

vector \(\mathbf{C} = (C_1, \dots,

Ck)\), where each \(Ci\) is an

integer in the range \(\{0, \dots,

p_i-1\}\).

Encoding Correspondence: By the construction of

the functorial framework (Section III.C) and the operational protocols

(Section IV.B), the physical encoding ensures a direct correspondence

\(Ci = ri\) for all channels \(i\). Therefore, the map from an integer to

its physical representation, \(N \mapsto

\mathbf{C}\), is also a bijection.

Unambiguous Representation: A bijective mapping

means that the Chern vector \(\mathbf{C}\) is a complete and unambiguous

representation of the integer \(N\).

There is a one-to-one correspondence, ensuring no information is lost.

This obviates any need for algorithmic reconstruction.

Direct Measurement: The multi-terminal

measurement protocol (Section IV.C) directly yields the integer

components of the Chern vector, \(\{C_1,

\dots, C_k\}\), via a set of independent, quantized Hall

conductance measurements.

Conclusion: The final output of the physical

system is the vector \(\mathbf{C}\),

which is demonstrably equivalent to the RNS representation of \(N\). The computational problem is solved

entirely by the physical process, and the result is directly readable

from the hardware.

Edge Case Analysis: The encoding is valid across

the entire dynamic range. For \(N =

0\), the system is in the vacuum state with \(\mathbf{C} = (0, \dots, 0)\). For \(N = M-1\), the system is in the state \(\mathbf{C} = (p1-1, \dots, pk-1)\). Both

are valid and stable ground states.

Robustness and Error Detection: Any error that

alters a single component \(C_i\)

results in a vector that does not correspond to a valid integer within

the intended computational range, enabling immediate error

detection.

Error Correction Potential: By choosing moduli

such that the total dynamic range \(M\)

exceeds the required range for a given problem, the redundant space can

be used to encode error-correcting information, enabling the correction

of single-channel errors.

Formal Measurement Map: The measurement process

can be formalized as a linear isomorphism \(\mathcal{M}: \mathbf{C} \mapsto

\{G_{xy}^{(i)}\}\), which maps the Chern vector space to the

conductance vector space, completing the formal chain from abstract

integer to physical observable.

References

Cheon, J. H., Kim, A., Kim, M., & Song, Y. (2017). Homomorphic

Encryption for Arithmetic of Approximate Numbers. In *Advances in

Cryptology–ASIACRYPT 2017* (pp. 409-437). Springer.

https://doi.org/10.1007/978-3-319-70697-9_15

Goguen, J. A., & Burstall, R. M. (1984). Some fundamental

algebraic tools for the semantics of computation. Part 1: Compositions

and refinements. Theoretical Computer Science, 31,

175-213. https://doi.org/10.1016/0304-3975(84)90004-0

Horowitz, M. (2014). Computing’s Energy Problem (and what we can do

about it). In *IEEE International Solid-State Circuits Conference

Digest of Technical Papers (ISSCC)* (pp. 10-14).

https://doi.org/10.1109/ISSCC.2014.6757323

Horsman, C., Kendon, V., Stepney, S., & Munro, W. J. (2014). When

does a physical system compute? *Proceedings of the Royal Society A:

Mathematical, Physical and Engineering Sciences*,

470(2171), 20140152. https://doi.org/10.1098/rspa.2014.0152

Kitaev, A. Y. (2003). Fault-tolerant quantum computation by anyons.

Annals of Physics, 303(1), 2-30.

https://doi.org/10.1016/S0003-4916(02)00018-0

Lloyd, S., & Pagels, H. (1988). Complexity as Thermodynamic

Depth. Annals of Physics, 188(1), 186-213.

https://doi.org/10.1016/0003-4916(88)90094-2

MacLennan, B. J. (2004). Natural Computation and Non-Turing Models of

Computation. Theoretical Computer Science, 317(1-3),

115-145. https://doi.org/10.1016/j.tcs.2003.12.006

Müller, V. C., & Hoffmann, M. (2017). What Is Morphological

Computation? On How the Body Contributes to Cognition and Control.

Artificial Life, 23(1), 1-24.

https://doi.org/10.1162/ARTLa00219

Nayak, C., Simon, S. H., Stern, A., Freedman, M., & Das Sarma, S.

(2008). Non-Abelian anyons and topological quantum computation.

Reviews of Modern Physics, 80(3), 1083-1159.

https://doi.org/10.1103/RevModPhys.80.1083

Pfeifer, R., & Bongard, J. (2007). *How the Body Shapes the

Way We Think: A New View of Intelligence*. MIT Press.

Prigogine, I. (1967). Structure, Dissipation and Life. In

Theoretical Physics and Biology (pp. 23-52). North Holland.

Sharpe, A. L., Fox, E. J., Barnard, A. W., Finney, J., Watanabe, K.,

Taniguchi, T., Kastner, M. A., & Goldhaber-Gordon, D. (2019).

Emergent ferromagnetism near three-quarters filling in twisted bilayer

graphene. Science, 365(6453), 605-608.

https://doi.org/10.1126/science.aaw3780

Siegelmann, H. T. (1999). Neural and Super-Turing Computing.

Minds and Machines, 13, 103-114.

https://doi.org/10.1023/A:1022902524528

Szabo, N. S., & Tanaka, R. I. (1967). *Residue Arithmetic and

Its Applications to Computer Technology*. McGraw-Hill.

Wen, X.-G. (2002). Quantum orders in an exact soluble model.

Physical Review B, 65(16), 165113.

https://doi.org/10.1103/PhysRevB.65.165113

Xie, Y., Pierce, A. T., Park, J. M., Parker, D. E., Khalaf, E.,

Ledwith, P., Lee, Y., Chen, S., Călugăru, D., Andrei, B. A., &

Yacoby, A. (2021). Fractional Chern insulators in magic-angle twisted

bilayer graphene. Nature, 597(7874), 38–43.

https://doi.org/10.1038/s41586-021-03842-4