arXiv subjects

Davide Nuzzi

Publications and source records attributed to Davide Nuzzi.

11 recordsLinked to original sources

arXiv2026

Grounded world models in biological organisms and future embodied AI

Recent advances in generative and embodied AI have been driven by large-scale predictive learning over multimodal data. However, the resulting systems remain largely based on passive training regimes where linguistic regularities create the scaffold onto which information from other modalities is attached. Conversely, neuroscience and cognitive science suggest that biological intelligence is organized in the opposite way, where grounded world models acquired through interaction with the environment provide the semantic scaffold to which language is attached. Here, we illustrate five examples of neural circuits supporting grounded world modelling, which underlie navigation in physical and conceptual spaces, affordance-based perception and interaction with objects, active perception and exploratory learning, allostatic control and emotion, and the distinction between self- and world-generated outcomes. These examples highlight several features largely missing from current embodied AI, including the role of intrinsic dynamics as a foundation for learning, the centrality of action in aligning these dynamics with the external world, the prominence of autonomous experience and open-ended learning over passive assimilation of externally provided data, and the fact that early predictive and control mechanisms scaffold higher cognitive abilities such as reasoning, conceptual navigation, planning, imagination, understanding others' minds, and communication. Finally, we discuss whether and how principles derived from biological systems may inform future embodied AI, including training regimes based on social interaction to construct world models that are not only grounded but also socially shared and aligned with human norms and values.

q-bio.NC↗

arXiv2025

A Rate-Distortion Perspective on the Emergence of Number Sense in Unsupervised Generative Models

Number sense is a core cognitive ability supporting various adaptive behaviors and is foundational for mathematical learning. Here, we study its emergence in unsupervised generative models through the lens of rate-distortion theory (RDT), a normative framework for understanding information processing under limited resources. We train $\beta$-Variational Autoencoders -- which embody key formal principles of RDT -- on synthetic images containing varying numbers of items, as commonly used in numerosity perception research. We systematically vary the encoding capacity and assess the models' sensitivity to numerosity and the robustness of the emergent numerical representations through a comprehensive set of analyses, including numerosity estimation and discrimination tasks, latent-space analysis, generative capabilities and generalization to novel stimuli. In line with RDT, we find that behavioral performance in numerosity perception and the ability to extract numerosity unconfounded by non-numerical visual features scale with encoding capacity according to a power law. At high capacity, the unsupervised model develops a robust neural code for numerical information, with performance closely approximating a supervised model explicitly trained for visual enumeration. It exhibits strong generative abilities and generalizes well to novel images, whereas at low capacity, the model shows marked deficits in numerosity perception and representation. Finally, comparison with human data shows that models trained at intermediate capacity levels span the full range of human behavioral performance while still developing a robust emergent numerical code. In sum, our results show that unsupervised generative models can develop a number sense and demonstrate that rate-distortion theory provides a powerful information-theoretic framework for understanding how capacity constraints shape numerosity perception.

q-bio.NC↗

arXiv2025

What the flock knows that the birds do not: exploring the emergence of joint agency in multi-agent active inference

Collective behavior pervades biological systems, from flocks of birds to neural assemblies and human societies. Yet, how such collectives acquire functional properties -- such as joint agency or knowledge -- that transcend those of their individual components remains an open question. Here, we combine active inference and information-theoretic analyses to explore how a minimal system of interacting agents can give rise to joint agency and collective knowledge. We model flocking dynamics using multiple active inference agents, each minimizing its own free energy while coupling reciprocally with its neighbors. We show that as agents self-organize, their interactions define higher-order statistical boundaries (Markov blankets) enclosing a ``flock'' that can be treated as an emergent agent with its own sensory, active, and internal states. When exposed to external perturbations (a ``predator''), the flock exhibits faster, coordinated responses than individual agents, reflecting collective sensitivity to environmental change. Crucially, analyses of synergistic information reveal that the flock encodes information about the predator's location that is not accessible to every individual bird, demonstrating implicit collective knowledge. Together, these results show how informational coupling among active inference agents can generate new levels of autonomy and inference, providing a framework for understanding the emergence of (implicit) collective knowledge and joint agency.

nlin.AO↗

arXiv2025

Structuring the Environment Nudges Participants Toward Hierarchical Over Shortest Path Planning

Effective planning is crucial for navigating complex environments and achieving goals efficiently. In this study, we investigated how environmental structure influences the selection of planning strategies. Forty-two participants navigated a space station to collect colored spheres, with environments either structured (spheres grouped by color) or unstructured (spheres scattered randomly). We tested three types of plans: hierarchical (grouping spheres by color), shortest path (minimizing travel distance), and neutral (none of the above). By manipulating environmental structure, we were able to nudge participants toward a preference for hierarchical planning in structured environments, while shortest path plans were favored in unstructured environments. A mismatch between self-reported preferences and actual choices indicated that participants often adopted implicit strategies, unaware of their decision-making processes. These findings highlight the powerful effect of environmental cues on planning and suggest that even subtle changes in structure can guide the selection of planning strategies.

q-bio.NC↗

arXiv2024

Human foraging strategies flexibly adapt to resource distribution and time constraints

Foraging is a crucial activity, yet the extent to which humans employ flexible versus rigid strategies remains unclear. This study investigates how individuals adapt their foraging strategies in response to resource distribution and foraging time constraints. For this, we designed a video-game-like foraging task that requires participants to navigate a four-areas environment to collect coins from treasure boxes within a limited time. This task engages multiple cognitive abilities, such as navigation, learning, and memorization of treasure box locations. Findings indicate that participants adjust their foraging strategies -- encompassing both stay-or-leave decisions, such as the number of boxes opened in initial areas and behavioral aspects, such as the time to navigate from box to box -- depending on both resource distribution and foraging time. Additionally, they improved their performance over time as an effect of both enhanced navigation skills and adaptation of foraging strategies. Finally, participants' performance was initially distant from the reward-maximizing performance of optimal agents due to the learning process humans undergo; however, it approximated the optimal agent's performance towards the end of the task, without fully reaching it. These results highlight the flexibility of human foraging behavior and underscore the importance of employing optimality models and ecologically rich scenarios to study foraging.

q-bio.NC↗

arXiv2023

Full-magnetic implementation of a classical Toffoli gate

The Toffoli gate is the essential ingredient for reversible computing, an energy efficient classical computational paradigm that evades the energy dissipation resulting from Landauer's principle. In this paper we analyze different setups to realize a magnetic implementation of the Toffoli gate using three interacting classical spins, each one embodying one of the three bits needed for the Toffoli gate. In our scheme, different control-spins configurations produce an effective field capable of conditionally flipping the target spin. We study what are the experimental requirements for the realization of our scheme, focusing on the degree of local control, the ability to dynamically switch the spin-spin interactions, and the required single-spin anisotropies to make the classical spin stable, showing that these are compatible with current technology.

quant-ph↗

arXiv2022

Gradients of O-information: low-order descriptors of high-order dependencies

O-information is an information-theoretic metric that captures the overall balance between redundant and synergistic information shared by groups of three or more variables. To complement the global assessment provided by this metric, here we propose the gradients of the O-information as low-order descriptors that can characterise how high-order effects are localised across a system of interest. We illustrate the capabilities of the proposed framework by revealing the role of specific spins in Ising models with frustration, and on practical data analysis on US macroeconomic data. Our theoretical and empirical analyses demonstrate the potential of these gradients to highlight the contribution of variables in forming high-order informational circuits

cs.IT↗

arXiv2021

Single-qubit remote manipulation by magnetic solitons

Magnetic solitons can constitute a means for manipulating qubits from a distance. This would overcome the necessity of directly applying selective magnetic fields, which is unfeasible in the case of a matrix of qubits embedded in a solid-state quantum device. If the latter contained one-dimensional Heisenberg spin chains coupled to each qubit, one can originate a soliton in a selected chain by applying a time-dependent field at one end of it, far from the qubits. The generation of realistic solitons has been simulated. When a suitable soliton passes by, the coupled qubit undergoes nontrivial operations, even in the presence of moderate thermal noise.

quant-ph↗

arXiv2020

Synergistic information in a dynamical model implemented on the human structural connectome reveals spatially distinct associations with age

We implement the dynamical Ising model on the large scale architecture of white matter connections of healthy subjects in the age range 4-85 years, and analyze the dynamics in terms of the synergy, a quantity measuring the extent to which the joint state of pairs of variables is projected onto the dynamics of a target one. We find that the amount of synergy in explaining the dynamics of the hubs of the structural connectivity (in terms of degree strength) peaks before the critical temperature, and can thus be considered as a precursor of a critical transition. Conversely the greatest amount of synergy goes into explaining the dynamics of more central nodes. We also find that the aging of the structural connectivity is associated to significant changes in the simulated dynamics: there are brain regions whose synergy decreases with age, in particular the frontal pole, the Subcallosal area and the Supplementary Motor area; these areas could then be more likely to show a decline in terms of the capability to perform higher order computation (if structural connectivity was the sole variable). On the other hand, several regions in the temporal cortex show a positive correlation with age in the first 30 years of life, i.e. during brain maturation.

q-bio.NC↗

arXiv2017

Quantum correlations between distant qubits conveyed by large-$S$ spin chains

We consider two distant spin-$\frac{1}{2}$ particles (or qubits) and a number of interacting objects, all with the same value $S\gg1$ of their respective spin, distributed on a one-dimensional lattice (or large-$S$ spin chain). The quantum states of the chain are constructed by linearly combining tensor products of single-spin coherent states, whose evolution is determined accordingly, i.e., via classical-like equations of motions. We show that the quantum superposition of the above product states resulting from a local interaction between the first qubit and one spin of the chain evolves so that the second qubit, after having itself interacted with another spin of the chain, can be entangled with the first qubit. Obtaining such outcome does not imply imposing constraints on the length of the chain or the distance between the qubits, which demonstrates the possibility of generating quantum correlations at a distance by means of a macroscopic system, as far as local interactions with just a few of its components are feasible.

quant-ph↗

arXiv2015

Getting through to a qubit by magnetic solitons

We propose a method for acting on the spin state of a spin-1/2 localized particle, or `qubit', by means of a magnetic signal effectively generated by the nearby transit of a magnetic soliton, there conveyed through a transmission line. We first introduce the specific magnetic soliton of which we will make use, and briefly review the properties that make it apt to represent a signal. We then show that a Heisenberg spin chain can serve as transmission line, and propose a method for injecting a soliton into the chain by acting just on one of its ends. We finally demonstrate that the resulting magnetic pulse can indeed cause, just passing by the spin-1/2 localized particle embodying the qubit, a permanent change in its spin state, thus realizing the possibility of getting through to a single, localized qubit, and manipulate its state. A thorough analysis of how the overall dynamical system operates depending on the setting of its parameters demonstrates that fine tuning is not necessary as it exists an extended region in the parameters space that corresponds to an effective functioning. Moreover, we show that possible noise on the transmission line does not invalidate the scheme.

cond-mat.mtrl-sci↗