arXiv subjects

Kazuya Horibe

Publications and source records attributed to Kazuya Horibe.

11 recordsLinked to original sources

arXiv2026

Emergence of Reputation-Based Cooperation in LLM Agents

Can cooperation among large language model (LLM) agents be evolutionarily stable against free-rider invasion? We study an indirect reciprocity donation game where LLM agents observe behavioral traces and donate on a continuous scale. Strategies, represented as natural language prompts, evolve through cultural transmission across generations. Across four LLM backends, robustness to free-rider invasion varies by more than an order of magnitude. The strongest predictor of this robustness is opponent endowment sensitivity, the degree to which agents discriminate between cooperative and uncooperative opponents, operationalizing the classical Image Scoring mechanism. By contrast, adherence to the Leading-Eight L1 norm does not predict robustness. Robustness depends on defector exclusion: while both cooperator reward and defector punishment vary across models, only the stringency of defector exclusion predicts resistance to free-rider invasion. These findings reveal that LLM agents are confined to Image Scoring-like discrimination and fail to develop the more robust Leading-Eight norms, highlighting a fundamental vulnerability in culturally evolved LLM cooperation and motivating bottom-up approaches to norm construction.

cs.MA↗

arXiv2026

Scale-Dependent Collective Adaptation in Self-Amending LLM Societies: A Cross-Family Study of Emergent Governance

We study group decision-making in artificial societies where the rules of play are themselves subject to collective amendment. Using the self-amending game Nomic, we compare multiple scales across two LLM families and find that collective adaptation does not improve monotonically with model size. Instead, both families exhibit a narrow mid-scale regime that supports sustained rule adoption, diverse amendments, and balanced consensus. Smaller models tend to remain rule-inert, whereas larger models often converge on restrictive voting patterns, and heterogeneous mixed-size groups collapse into veto-driven gridlock. These cross-scale contrasts persist under temperature perturbations and under a shift from unanimity to majority voting, although latent-state structure varies by family and scale. Hidden-state divergence alone does not explain collective performance: high representational divergence can coincide with poor behavioural outcomes. Linear probes reveal regime-selective coupling between latent vote-predictive signals and collective behaviour, but decodability is necessary rather than sufficient for adaptive play. Overall, the recurring regularity is non-monotonicity, not the particular scale at which the optimum appears. Self-amending games therefore provide a controlled testbed for studying collective adaptation in artificial societies beyond raw model scale.

nlin.AO↗

arXiv2025

Agency at the Interface: Distinguishing Teleological from Structural Self-Organization via Internal Coarse-Graining and Downward Causation

Agency is widely characterized as the capacity of a system to regulate its internal states toward self-generated goals, yet characterizing the functional basis of this autonomy requires a distinction between the system's own organization and an observer's interpretation. In this paper, we ground this capacity functionally in the generative process of intrinsic dynamics, characterized as a self-referential loop generated by internal coarse-graining and downward causation. This process allows the system to autonomously compress microscopic states into macroscopic variables that subsequently constrain microscopic temporal evolution. By distinguishing the intrinsic dynamics of the system from the external coarse-graining of an observer's interpretation, we define agency through the dynamics of the predictive gap. This gap constitutes the internal divergence between the system's anticipation and its realization, as well as the limited reducibility of the system's generated constraints within the observer's interpretive model. This framework outlines a spectrum of agency that distinguishes biological systems, whose teleological constraints are generated through their own regulatory dynamics, from artificial systems, whose organization is designed to satisfy externally specified objectives. Finally, we extend this interface to the social domain, proposing that these generative divergences underpin participatory sense-making and emergent coordination.

nlin.AO↗

arXiv2025

Evolvable Chemotons: Toward the Integration of Autonomy and Evolution

In this study, we provide a relatively simple simulation framework for constructing artificial life (ALife) with both autonomous and evolutionary aspects by extending chemoton model. While the original chemoton incorporates metabolism, membrane, and genetic templates, it lacks a mechanism for phenotypic variation, preventing true evolutionary dynamics. To address this, we introduced a genotype-phenotype coupling by linking templates to a second autocatalytic cycle, enabling mutations to affect phenotype and be subject to selection. Using a genetic algorithm, we simulated populations of chemotons over generations. Results showed that chemotons without access to the new cycle remained in a stable but complexity-limited regime, while lineages acquiring the additional metabolic set evolved longer templates. These findings demonstrate that even simple replicator systems can achieve primitive evolvability, highlighting structural thresholds and rare innovations as key drivers. Our framework provides a tractable model for exploring autonomy and evolution in ALife.

q-bio.PE↗

arXiv2024

Emergence of Implicit World Models from Mortal Agents

We discuss the possibility of world models and active exploration as emergent properties of open-ended behavior optimization in autonomous agents. In discussing the source of the open-endedness of living things, we start from the perspective of biological systems as understood by the mechanistic approach of theoretical biology and artificial life. From this perspective, we discuss the potential of homeostasis in particular as an open-ended objective for autonomous agents and as a general, integrative extrinsic motivation. We then discuss the possibility of implicitly acquiring a world model and active exploration through the internal dynamics of a network, and a hypothetical architecture for this, by combining meta-reinforcement learning, which assumes domain adaptation as a system that achieves robust homeostasis.

cs.NE↗

arXiv2023

Exploring the Adaptive Behaviors of Particle Lenia: A Perturbation-Response Analysis for Computational Agency

A firm cognitive subject or ``individual'' is presupposed for the emergence of mind. However, with the development of recent information technology, the ``individual'' has become more dispersed in society and the cognitive subject has become increasingly unstable and adaptive, necessitating an update in our understanding of the ``individual''. Autopoiesis serves as a model of the cognitive subject, which is unstable and requires effort to maintain itself to adapt to the environment. In this study, we evaluated adaptivity for a highly extensible multi-particle system model Particle Lenia through the response perturbation. As a result, we found that Particle Lenia has a particle configuration that is both temporally unstable and has multiple stable states. This result suggests that Particle Lenia can express adaptive characteristics and is expected to be used as a computational model toward building an autopoietic cognitive agent.

nlin.AO↗

arXiv2022

Severe Damage Recovery in Evolving Soft Robots through Differentiable Programming

Biological systems are very robust to morphological damage, but artificial systems (robots) are currently not. In this paper we present a system based on neural cellular automata, in which locomoting robots are evolved and then given the ability to regenerate their morphology from damage through gradient-based training. Our approach thus combines the benefits of evolution to discover a wide range of different robot morphologies, with the efficiency of supervised training for robustness through differentiable update rules. The resulting neural cellular automata are able to grow virtual robots capable of regaining more than 80\% of their functionality, even after severe types of morphological damage.

cs.NE↗

arXiv2022

Investigating the impact of free energy based behavior on human in human-agent interaction

Humans communicate non-verbally by sharing physical rhythms, such as nodding and gestures, to involve each other. This sharing of physicality creates a sense of unity and makes humans feel involved with others. In this paper, we developed a new body motion generation system based on the free-energy principle (FEP), which not only responds passively but also prompts human actions. The proposed system consists of two modules, the sampling module, and the motion selection module. We conducted a subjective experiment to evaluate the "feeling of interacting with the agent" of the FEP based behavior. The results suggested that FEP based behaviors show more "feeling of interacting with the agent". Furthermore, we confirmed that the agent's gestures elicited subject gestures. This result not only reinforces the impression of feeling interaction but could also realization of agents that encourage people to change their behavior.

cs.HC↗

arXiv2021

Regenerating Soft Robots through Neural Cellular Automata

Morphological regeneration is an important feature that highlights the environmental adaptive capacity of biological systems. Lack of this regenerative capacity significantly limits the resilience of machines and the environments they can operate in. To aid in addressing this gap, we develop an approach for simulated soft robots to regrow parts of their morphology when being damaged. Although numerical simulations using soft robots have played an important role in their design, evolving soft robots with regenerative capabilities have so far received comparable little attention. Here we propose a model for soft robots that regenerate through a neural cellular automata. Importantly, this approach only relies on local cell information to regrow damaged components, opening interesting possibilities for physical regenerable soft robots in the future. Our approach allows simulated soft robots that are damaged to partially regenerate their original morphology through local cell interactions alone and regain some of their ability to locomote. These results take a step towards equipping artificial systems with regenerative capacities and could potentially allow for more robust operations in a variety of situations and environments. The code for the experiments in this paper is available at: \url{github.com/KazuyaHoribe/RegeneratingSoftRobots}.

cs.NE↗

arXiv2019

Cell Motion Alignment as Polarity Memory Effect

The clarification of the motion alignment mechanism in collective cell migration is an important issue commonly in physics and biology. In analogy with the self-propelled disk, the polarity memory effect of eukaryotic cell is a fundamental candidate for this alignment mechanism. In the present paper, we theoretically examine the polarity memory effect for the motion alignment of cells on the basis of the cellular Potts model. We show that the polarity memory effect can align motion of cells. We also find that the polarity memory effect emerges for the persistent length of cell trajectories longer than average cell-cell distance.

q-bio.CB↗

arXiv2019

Curved surface geometry-induced topological change of an excitable planar wave

On the curved surfaces of living and nonliving materials, planar excitable waves frequently exhibit directional change and subsequently undergo a topological change; that is, a series of wave dynamics from fusion, annihilation to splitting. Theoretical studies have shown that excitable planar stable waves change their topology significantly depending on the initial conditions on flat surfaces, whereas the directional-change of the waves occurs based on the geometry of curved surfaces. However, it is not clear if the geometry of curved surfaces induces this topological change. In this study, we first show the curved surface geometry-induced topological changes in a planar stable wave by numerically solving an excitable reaction-diffusion equation on a bell-shaped surface. We determined two necessary conditions for inducing topological change: the characteristic length of the curved surface (i.e., height of the bell-shaped structure) should be larger than the width of the wave and than a threshold independent of the wave width. As for the geometrical mechanism of the latter, we found that a bifurcation of the globally minimum geodesics (i.e. minimal paths) on the curved surface leads to the topological change. These conditions imply that wave topology changes can be predicted on the basis of curved surfaces, whose structure is larger than the wave width.

nlin.PS↗