Institution profile

Università degli Studi di Napoli Federico II

Academic institutioneurope · it
Official website
Research library234linked papers
Opportunities0open roles
Selected work

Representative Papers

Entanglement-Based Artificial Topology: Neighboring Remote Network Nodes

Apr 24, 2024IEEE Open Journal of the Communications Society

To address the limitations of fixed physical topologies and restricted inter-network connectivity in quantum local area networks (QLANs), this paper proposes a dynamic network-layer topology reconfiguration method based on multipartite entanglement. By designing distributable multipartite entangled states and integrating local quantum operations with a novel entanglement routing protocol, the approach enables on-demand construction of “artificial topologies” and “artificial neighborhoods” across QLANs—overcoming functional bottlenecks inherent to bipartite entanglement in inter-network linking. This work constitutes the first systematic application of multipartite entanglement to quantum network topology engineering, supporting real-time, traffic-driven reconfiguration of artificial network structures. Experimental validation demonstrates significant improvements in logical connectivity between remote nodes, network flexibility, and entanglement resource utilization. The method establishes a new paradigm for scalable and engineerable quantum internet architectures.

9 citationsRead paper

Module checking of pushdown multi-agent systems

Mar 07, 2020International Conference on Principles of Knowledge Representation and Reasoning

This paper investigates the module-checking problem for pushdown multi-agent systems (PMS) against specifications in Alternating-Time Temporal Logic (ATL) and its extension ATL*. Leveraging a synthesis of game-theoretic semantics, fixed-point analysis of pushdown systems, and complexity-theoretic reductions, we establish—first time—the exact computational complexities: ATL module-checking is 2EXPTIME-complete, and ATL* module-checking is 4EXPTIME-complete. The latter represents one of the rare natural decidable problems with complexity strictly above triple-exponential time, markedly exceeding both pushdown CTL* module-checking (3EXPTIME-complete) and ATL* model-checking (3EXPTIME-complete). This exponential leap underscores the intrinsic complexity arising from the interplay between modular architecture and strategic interaction among agents. Our results provide fundamental theoretical limits and methodological foundations for high-assurance verification of multi-agent systems.

4 citationsRead paper

A Hybrid Model-based and Data-based Approach Developed for a Prosthetic Hand Wrist

Jan 13, 2026

This work proposes a hybrid controller integrating artificial neural networks (ANN) and sliding mode control (SMC) to enhance the dynamic response and reduce computational load in tendon-driven soft continuum prosthetic wrists. For the first time, this approach is applied to such a system, leveraging a kinematic and dynamic model based on the piecewise constant curvature assumption. The ANN estimates wrist bending angles in real time, while the SMC precisely regulates tendon tension. Comprehensive simulations and experiments conducted on the PRISMA HAND II platform demonstrate that the proposed method achieves high tracking accuracy while significantly improving response speed and reducing computational overhead. The results highlight superior dynamic performance and robustness compared to existing control strategies.

2 citationsRead paper

PerspAct: Enhancing LLM Situated Collaboration Skills through Perspective Taking and Active Vision

Nov 11, 2025

Current large language models (LLMs) and multimodal models exhibit limited perspective-taking capabilities in multi-agent collaboration, hindering accurate modeling of subjective agent perceptions and multi-observer environments. To address this, we propose PerspAct—a novel method that integrates active visual exploration with the ReAct reasoning framework for the first time. PerspAct explicitly samples and models diverse agent-centric perspectives, enabling dynamic comprehension of hierarchical perspective complexity in an extended Director task. Built upon multimodal LLMs, it leverages prompt engineering and explicit state representation. We systematically evaluate PerspAct across seven progressively complex scenarios. Experiments demonstrate significant improvements in both coreference resolution and collaborative task accuracy, validating the efficacy of jointly modeling active perception and perspective understanding. Our work establishes a new paradigm for situational awareness in multi-agent settings.

1 citations1 influentialRead paper
Recent publications

Latest Papers