-
From bordisms of three-manifolds to domain walls between topological orders
Authors:
Yu Leon Liu,
Dalton A R Sakthivadivel
Abstract:
We study a correspondence between spin three-manifolds and bosonic abelian topological orders. Let $N$ be a spin three-manifold. We can define a $(2+1)$-dimensional topological order $\mathrm{TO}_N$ as follows: its anyons are the torsion elements in $H_1(N)$, the braiding of anyons is given by the linking form, and their topological spins are given by the quadratic refinement of the linking form o…
▽ More
We study a correspondence between spin three-manifolds and bosonic abelian topological orders. Let $N$ be a spin three-manifold. We can define a $(2+1)$-dimensional topological order $\mathrm{TO}_N$ as follows: its anyons are the torsion elements in $H_1(N)$, the braiding of anyons is given by the linking form, and their topological spins are given by the quadratic refinement of the linking form obtained from the spin structure. Under this correspondence, a surgery presentation of $N$ gives rise to a classical Chern--Simons description of the associated topological order $\mathrm{TO}_N$. We then extend the correspondence to spin bordisms between three-manifolds, and domain walls between topological orders. In particular, we construct a domain wall $\mathcal{D}_M$ between $\mathrm{TO}_N$ and $\mathrm{TO}_{N'}$, where $M$ is a spin bordism from $N$ to $N'$. This domain wall unfolds to a composition of a gapped boundary, obtained from anyon condensation, and a gapless Narain boundary CFT.
△ Less
Submitted 15 July, 2024;
originally announced July 2024.
-
A framework for the use of generative modelling in non-equilibrium statistical mechanics
Authors:
Maxwell J D Ramstead,
Dalton A R Sakthivadivel,
Karl J Friston
Abstract:
We discuss an approach to mathematically modelling systems made of objects that are coupled together, using generative models of the dependence relationships between states (or trajectories) of the things comprising such systems. This broad class includes open or non-equilibrium systems and is especially relevant to self-organising systems. The ensuing variational free energy principle (FEP) has c…
▽ More
We discuss an approach to mathematically modelling systems made of objects that are coupled together, using generative models of the dependence relationships between states (or trajectories) of the things comprising such systems. This broad class includes open or non-equilibrium systems and is especially relevant to self-organising systems. The ensuing variational free energy principle (FEP) has certain advantages over using random dynamical systems explicitly, notably, by being more tractable and offering a parsimonious explanation of why the joint system evolves in the way that it does, based on the properties of the coupling between system components. Using the FEP allows us to model the dynamics of an object as if it were a process of variational inference, because variational free energy (or surprisal) is a Lyapunov function for its dynamics. In short, we argue that using generative models to represent and track relations among subsystems leads us to a particular statistical theory of interacting systems. Conversely, this theory enables us to construct nested models that respect the known relations among subsystems. We point out that the fact that a physical object conforms to the FEP does not necessarily imply that this object performs inference in the literal sense; rather, it is a useful explanatory fiction which replaces the `explicit' dynamics of the object with an `implicit' flow on free energy gradients -- a fiction that may or may not be entertained by the object itself.
△ Less
Submitted 24 February, 2025; v1 submitted 17 June, 2024;
originally announced June 2024.
-
Functorial Statistical Physics: Feynman--Kac Formulae and Information Geometries
Authors:
Dalton A R Sakthivadivel
Abstract:
The main results of this paper comprise proofs of the following two related facts: (i) the Feynman--Kac formula is a functor $F_*$, namely, between a stochastic differential equation and a dynamical system on a statistical manifold, and (ii) a statistical manifold is a sheaf generated by this functor with a canonical gluing condition. Using a particular locality property for $F_*$, recognised from…
▽ More
The main results of this paper comprise proofs of the following two related facts: (i) the Feynman--Kac formula is a functor $F_*$, namely, between a stochastic differential equation and a dynamical system on a statistical manifold, and (ii) a statistical manifold is a sheaf generated by this functor with a canonical gluing condition. Using a particular locality property for $F_*$, recognised from functorial quantum field theory as a `sewing law,' we then extend our results to the Chapman--Kolmogorov equation {\it via} a time-dependent generalisation of the principle of maximum entropy. This yields a partial formalisation of a variational principle which takes us beyond Feynman--Kac measures driven by Wiener laws. Our construction offers a robust glimpse at a deeper theory which we argue re-imagines time-dependent statistical physics and information geometry alike.
△ Less
Submitted 27 December, 2022;
originally announced December 2022.
-
Path integrals, particular kinds, and strange things
Authors:
Karl Friston,
Lancelot Da Costa,
Dalton A. R. Sakthivadivel,
Conor Heins,
Grigorios A. Pavliotis,
Maxwell Ramstead,
Thomas Parr
Abstract:
This paper describes a path integral formulation of the free energy principle. The ensuing account expresses the paths or trajectories that a particle takes as it evolves over time. The main results are a method or principle of least action that can be used to emulate the behaviour of particles in open exchange with their external milieu. Particles are defined by a particular partition, in which i…
▽ More
This paper describes a path integral formulation of the free energy principle. The ensuing account expresses the paths or trajectories that a particle takes as it evolves over time. The main results are a method or principle of least action that can be used to emulate the behaviour of particles in open exchange with their external milieu. Particles are defined by a particular partition, in which internal states are individuated from external states by active and sensory blanket states. The variational principle at hand allows one to interpret internal dynamics - of certain kinds of particles - as inferring external states that are hidden behind blanket states. We consider different kinds of particles, and to what extent they can be imbued with an elementary form of inference or sentience. Specifically, we consider the distinction between dissipative and conservative particles, inert and active particles and, finally, ordinary and strange particles. Strange particles can be described as inferring their own actions, endowing them with apparent autonomy or agency. In short - of the kinds of particles afforded by a particular partition - strange kinds may be apt for describing sentient behaviour.
△ Less
Submitted 13 June, 2023; v1 submitted 23 October, 2022;
originally announced October 2022.
-
Weak Markov Blankets in High-Dimensional, Sparsely-Coupled Random Dynamical Systems
Authors:
Dalton A R Sakthivadivel
Abstract:
This paper formulates a notion of high-dimensional random dynamical systems that couple to another system, like an embedding environment, in such a way that each system engages in controlled exchange with the other system. Using the adiabatic theorem and asymptotic arguments about the behaviour of systems with many degrees of freedom, we show that this sort of controlled, but not strictly sparse,…
▽ More
This paper formulates a notion of high-dimensional random dynamical systems that couple to another system, like an embedding environment, in such a way that each system engages in controlled exchange with the other system. Using the adiabatic theorem and asymptotic arguments about the behaviour of systems with many degrees of freedom, we show that this sort of controlled, but not strictly sparse, coupling structure is ubiquitous in high-dimensional systems. In doing so we prove a conjecture of K Friston.
△ Less
Submitted 14 August, 2022; v1 submitted 15 July, 2022;
originally announced July 2022.
-
A Worked Example of the Bayesian Mechanics of Classical Objects
Authors:
Dalton A R Sakthivadivel
Abstract:
Bayesian mechanics is a new approach to studying the mathematics and physics of interacting stochastic processes. Here, we provide a worked example of a physical mechanics for classical objects, which derives from a simple application thereof. We summarise the current state of the art of Bayesian mechanics in doing so. We also give a sketch of its connections to classical chaos, owing to a particu…
▽ More
Bayesian mechanics is a new approach to studying the mathematics and physics of interacting stochastic processes. Here, we provide a worked example of a physical mechanics for classical objects, which derives from a simple application thereof. We summarise the current state of the art of Bayesian mechanics in doing so. We also give a sketch of its connections to classical chaos, owing to a particular $\mathcal{N}=2$ supersymmetry.
△ Less
Submitted 6 September, 2022; v1 submitted 26 June, 2022;
originally announced June 2022.
-
Towards a Geometry and Analysis for Bayesian Mechanics
Authors:
Dalton A R Sakthivadivel
Abstract:
In this paper, a simple case of Bayesian mechanics under the free energy principle is formulated in axiomatic terms. We argue that any dynamical system with constraints on its dynamics necessarily looks as though it is performing inference against these constraints, and that in a non-isolated system, such constraints imply external environmental variables embedding the system. Using aspects of cla…
▽ More
In this paper, a simple case of Bayesian mechanics under the free energy principle is formulated in axiomatic terms. We argue that any dynamical system with constraints on its dynamics necessarily looks as though it is performing inference against these constraints, and that in a non-isolated system, such constraints imply external environmental variables embedding the system. Using aspects of classical dynamical systems theory in statistical mechanics, we show that this inference is equivalent to a gradient ascent on the Shannon entropy functional, recovering an approximate Bayesian inference under a locally ergodic probability measure on the state space. We also use some geometric notions from dynamical systems theory$\unicode{x2014}$namely, that the constraints constitute a gauge degree of freedom$\unicode{x2014}$to elaborate on how the desire to stay self-organised can be read as a gauge force acting on the system. In doing so, a number of results of independent interest are given. Overall, we provide a related, but alternative, formalism to those driven purely by descriptions of random dynamical systems, and take a further step towards a comprehensive statement of the physics of self-organisation in formal mathematical language.
△ Less
Submitted 25 April, 2022;
originally announced April 2022.
-
Entropy-Maximising Diffusions Satisfy a Parallel Transport Law
Authors:
Dalton A R Sakthivadivel
Abstract:
We show that the principle of maximum entropy, a variational method appearing in statistical inference, statistical physics, and the analysis of stochastic dynamical systems, admits a geometric description from gauge theory. Using the connection on a principal $G$-bundle, the gradient flow maximising entropy is written in terms of constraint functions which interact with the dynamics of the probab…
▽ More
We show that the principle of maximum entropy, a variational method appearing in statistical inference, statistical physics, and the analysis of stochastic dynamical systems, admits a geometric description from gauge theory. Using the connection on a principal $G$-bundle, the gradient flow maximising entropy is written in terms of constraint functions which interact with the dynamics of the probabilistic degrees of freedom of a diffusion process. This allows us to describe the point of maximum entropy as parallel transport over the state space. In particular, it is proven that the solubility of the stationary Fokker--Planck equation corresponds to the existence of parallel transport in a particular associated vector bundle, extending classic results due to Jordan--Kinderlehrer--Otto and Markowich--Villani. A reinterpretation of splitting results in stochastic dynamical systems is also suggested. Beyond stochastic analysis, we are able to indicate a collection of geometric structures surrounding energy-based inference in statistics.
△ Less
Submitted 3 January, 2023; v1 submitted 15 March, 2022;
originally announced March 2022.
