arXiv subjects

Eric De Giuli

Publications and source records attributed to Eric De Giuli.

13 recordsLinked to original sources

arXiv2026

Entangled criticality and irreversibility in random Markov dynamics

We introduce a two-parameter ensemble of random discrete-time Markov models that simultaneously captures critical slowing down and broken detailed balance. Extending a previously studied heterogeneous Markov ensemble, we incorporate correlations between forward and backward transition rates through a single asymmetry parameter $γ$, while heterogeneity is controlled by $ε$. Using results from random matrix theory, we identify a critical locus $ε_c(γ,N)$ at which relaxation times diverge and spectral universality breaks down, in Markov models with $N$ states. We characterize the behavior of entropy production, predictive information, and relaxation dynamics across the ensemble, showing that many observables depend strongly on heterogeneity but only weakly on asymmetry, except near the symmetric limit. Applying maximum-likelihood inference to human fMRI and EEG data, we find that both modalities operate near the predicted critical locus and occupy a similar region of the $ε-γ$ plane, supporting a super-universality of human brain dynamics. While ensemble averages are well captured by the null model, empirical data exhibit substantially enhanced variability, indicating subject-specific structure beyond random expectations. Our results unify criticality and nonequilibrium measures within a single framework and clarify their intertwined role in the analysis of complex biological dynamics.

cond-mat.dis-nn↗

arXiv2026

Phase structure of the Random Language Model

Context-free grammars are minimal models of hierarchical structure in human language, generating structured text from recursive production rules. The Random Language Model (RLM) [De Giuli, PRL 2019], an ensemble of such grammars with random rule weights, exhibits a cross-over from gibberish-like output to structured text as a function of a "temperature", but the location and nature of this transition remained unclear. Here, we show that the RLM exhibits a hierarchy of phase transitions in a double-scaling limit where the grammar temperature $\tildeε_d \to 0$ and the number of hidden symbols $N \to \infty$ at fixed $x = \tildeε_d \log N$. By identifying the relation between RLM and the Random Energy Model, we identify a series of transitions where correlations between symbols emerge, single-symbol marginals become non-uniform, and rule use freezes in a glassy phase. A semi-annealed approximation yields nontrivial scaling laws for rule usage, entropy, and energy, consistent with Heaps' law and context-length scaling observed in large language models.

cond-mat.dis-nn↗

arXiv2026

Scaling limit of the Random Language Model

We develop a quantitative theory of the Random Language Model (RLM), an ensemble of stochastic context-free grammars, in a scaling limit where the number of hidden symbols $N \to \infty$ while the grammar temperature $\tildeε_d \to 0$ at fixed $x = {\tildeε}_d \log N$. In this limit, the model admits a controlled description based on a large-deviation principle over rule-usage patterns. A semi-annealed approximation maps the problem to a class of Random Energy Models with nontrivial combinatorics. We show that the RLM exhibits a condensation transition at a critical value $x_c=1/8$, below which rule usage concentrates and language statistics acquire a nontrivial dependence on corpus length. A second characteristic scale at $x=1/2$ marks the onset of entropy reduction from its maximal value. Across these regimes, we derive explicit scaling laws for the number of distinct rules, entropy, and related observables, identifying distinct scaling, saturation, and critical regimes controlled by the interplay of grammar size, corpus length, and temperature. The theory resolves previous ambiguities regarding the existence of a thermodynamic transition and explains the slow approach to the large-$N$ limit as a consequence of the dependence on $\log N$. It further provides a unified framework in which universal statistical properties of language emerge from typical realizations of generative grammars, with implications for both natural language statistics and the behavior of large language models.

cond-mat.dis-nn↗

arXiv2025

Noise equals control

Stochastic systems have a control-theoretic interpretation in which noise plays the role of control. In the weak-noise limit, relevant at low temperatures or in large populations, this leads to a precise mathematical mapping: the most probable trajectory between two states minimizes an action functional and corresponds to an optimal control strategy. In Langevin dynamics, the noise term itself serves as the control. For general Markov jump processes, such as chemical reaction networks or electronic circuits, we use the Doi-Peliti formalism to identify the `response' (or `momentum') field $π$ as the control variable. This resolves a long-standing interpretational problem in the field-theoretic description of stochastic systems: although $π$ evolves backward in time, it has a clear physical role as the control that steers the system along rare trajectories. This implies that Nature is constantly sampling control strategies. We illustrate the mapping on multistable chemical reaction networks, systems with unstable fixed points, and specifically on stochastic resonance and Brownian ratchets. The noise-control mapping justifies agential descriptions of these phenomena, and builds intuition for otherwise puzzling phenomena of stochastic systems: why probabilities are generically non-smooth functions of state out of thermal equilibrium; why biological mechanisms can work better in the presence of noise; and how agential behavior emerges naturally without recourse to mysticism.

q-bio.MN↗

arXiv2024

Fermionic theory of nonequilibrium steady states

As the quantification of metabolism, nonequilibrium steady states play a central role in living matter, but are beyond the purview of equilibrium statistical mechanics. Here we develop a fermionic theory of nonequilibrium steady states in continuous-time Markovian systems, generalizing Boltzmann-Gibbs statistical mechanics to this case. The response to an arbitrary perturbation is computed, and simplified in canonical cases. Beyond response, we consider ensembles of nonequilibrium steady states and show that a general class of ensembles is described by a 2D statistical field theory with infinitesimally broken supersymmetry, which may form the basis of nontrivial solvable models of nonequilibrium steady states.

cond-mat.stat-mech↗

arXiv2024

Robustness of the Random Language Model

The Random Language Model (De Giuli 2019) is an ensemble of stochastic context-free grammars, quantifying the syntax of human and computer languages. The model suggests a simple picture of first language learning as a type of annealing in the vast space of potential languages. In its simplest formulation, it implies a single continuous transition to grammatical syntax, at which the symmetry among potential words and categories is spontaneously broken. Here this picture is scrutinized by considering its robustness against extensions of the original model, and trajectories through parameter space different from those originally considered. It is shown here that (i) the scenario is robust to explicit symmetry breaking, an inevitable component of learning in the real world; and (ii) the transition to grammatical syntax can be encountered by fixing the deep (hidden) structure while varying the surface (observable) properties. It is also argued that the transition becomes a sharp thermodynamic transition in an idealized limit. Moreover, comparison with human data on the clustering coefficient of syntax networks suggests that the observed transition is equivalent to that normally experienced by children at age 24 months. The results are discussed in light of theory of first-language acquisition in linguistics, and recent successes in machine learning.

cond-mat.dis-nn↗

arXiv2024

Universal Slow Dynamics of Chemical Reaction Networks

Understanding the emergent behavior of chemical reaction networks (CRNs) is a fundamental aspect of biology and its origin from inanimate matter. A closed CRN monotonically tends to thermal equilibrium, but when it is opened to external reservoirs, a range of behaviors is possible, including transition to a new equilibrium state, a non-equilibrium state, or indefinite growth. This study shows that slowly driven CRNs are governed by the conserved quantities of the closed system, which are generally far fewer in number than the species. Considering both deterministic and stochastic dynamics, a universal slow dynamics equation is derived with singular perturbation methods, and is shown to be thermodynamically consistent. The slow dynamics is highly robust against microscopic details of the network, which may be unknown in practical situations. In particular, non-equilibrium states of realistic large CRNs can be sought without knowledge of bulk reaction rates. The framework is successfully tested against a suite of networks of increasing complexity and argued to be relevant in the treatment of open CRNs as chemical machines.

q-bio.MN↗

arXiv2022

Dynamical mean-field theory: from ecosystems to reaction networks

Both natural ecosystems and biochemical reaction networks involve populations of heterogeneous agents whose cooperative and competitive interactions lead to a rich dynamics of species' abundances, albeit at vastly different scales. The maintenance of diversity in large ecosystems is a longstanding puzzle, towards which recent progress has been made by the derivation of dynamical mean-field theories of random models. In particular, it has recently been shown that these random models have a chaotic phase in which abundances display wild fluctuations. When modest spatial structure is included, these fluctuations are stabilized and diversity is maintained. If and how these phenomena have parallels in biochemical reaction networks is currently unknown. Making this connection is of interest since life requires cooperation among a large number of molecular species. In this work, we find a reaction network whose large-scale behavior recovers the random Lotka-Volterra model recently considered in theoretical ecology. We clarify the assumptions necessary to derive its large-scale description, and reveal the underlying assumptions made on the noise to recover previous dynamical mean-field theories. Then, we show how local detailed balance and the positivity of reaction rates, which are key physical requirements of chemical reaction networks, provide obstructions towards the construction of an associated dynamical mean-field theory of biochemical reaction networks. Finally, we outline prospects and challenges for the future.

cond-mat.stat-mech↗

arXiv2022

Emergence of order in random languages

We consider languages generated by weighted context-free grammars. It is shown that the behaviour of large texts is controlled by saddle-point equations for an appropriate generating function. We then consider ensembles of grammars, in particular the Random Language Model of E. DeGiuli, Phys. Rev. Lett., 122, 128301, 2019. This model is solved in the replica-symmetric ansatz, which is valid in the high-temperature, disordered phase. It is shown that in the phase in which languages carry information, the replica symmetry must be broken.

cond-mat.dis-nn↗

arXiv2021

Random quench predicts universal properties of amorphous solids

Amorphous solids display numerous universal features in their mechanics, structure, and response. Current models assume heterogeneity in mesoscale elastic properties, but require fine-tuning in order to quantitatively explain vibrational properties. A complete model should derive the magnitude and character of elastic heterogeneity from an initially structureless medium, through a model of the quenching process during which the temperature is rapidly lowered and the solid is formed. Here we propose a field-theoretic model of a quench, and compute structural, mechanical, and vibrational observables in arbitrary dimension $d$. This allows us to relate the properties of the amorphous solid to those of the initial medium, and to those of the quench. We show that previous mean-field results are subsumed by our analysis and unify spatial fluctuations of elastic moduli, long-range correlations of inherent state stress, universal vibrational anomalies, and localized modes into one picture.

cond-mat.dis-nn↗

arXiv2021

Breakdown of random matrix universality in Markov models

Biological systems need to react to stimuli over a broad spectrum of timescales. If and how this ability can emerge without external fine-tuning is a puzzle. We consider here this problem in discrete Markovian systems, where we can leverage results from random matrix theory. Indeed, generic large transition matrices are governed by universal results, which predict the absence of long timescales unless fine-tuned. We consider an ensemble of transition matrices and motivate a temperature-like variable that controls the dynamic range of matrix elements, which we show plays a crucial role in the applicability of the large matrix limit: as the dynamic range increases, a phase transition occurs whereby the random matrix theory result is avoided, and long relaxation times ensue, in the entire `ordered' phase. We furthermore show that this phase transition is accompanied by a drop in the entropy rate and a peak in complexity, as measured by predictive information (Bialek, Nemenman, Tishby Neural Computation 13(21) 2001). Extending the Markov model to a Hidden Markov model (HMM), we show that observable sequences inherit properties of the hidden sequences, allowing HMMs to be understood in terms of more accessible Markov models. We then apply our findings to fMRI data from 820 human subjects scanned at wakeful rest. We show that the data can be quantitatively understood in terms of the random model, and that brain activity lies close to the phase transition when engaged in unconstrained, task-free cognition -- supporting the brain criticality hypothesis in this context.

cond-mat.dis-nn↗

arXiv2021

Glassy gravity

Euclidean quantum gravity is reconsidered, in the conformal mode approximation. Applying techniques from glass theory, we argue that the Euclidean partition function hides metastable states, which can be counted. This may reconcile conflicting results on the uniqueness of the de Sitter vacuum, and may be relevant to the cosmological constant problem.

gr-qc↗

arXiv2020

Renormalization of elastic quadrupoles in amorphous solids

Plasticity in amorphous solids is mediated by localized quadrupolar instabilities, but the mechanism by which an amorphous solid eventually fails or melts is debated. In this work we argue that these phenomena can be investigated in the model problem of an elastic continuum with quadrupolar defects, at finite temperature. This problem is posed and the collective behavior of the defects is analytically investigated. Using both renormalization group and field-theoretic techniques, it is found that the model has a yielding/melting transition of spinodal type.

cond-mat.stat-mech↗