arXiv subjects

Gabriel Almeida

Publications and source records attributed to Gabriel Almeida.

8 recordsLinked to original sources

arXiv2026

Effectiveness of LLM-based Software Diversity for Reliability Improvement -- an Empirical Study

Software diversity has been extensively studied as a means of reducing the risk of common-mode failures. Classic work showed that the central issue is whether failures of diversely redundant components overlap in ways that limit the reliability gains. Traditional software diversity is costly to obtain, since it requires multiple implementations as well as the corresponding validation, maintenance, and deployment effort. Recent advances in Large Language Models (LLMs) may change this. LLMs enable inexpensive code generation: they produce many candidate implementations of the same specification quickly, across different models, decoding settings, and programming languages. This raises a natural question: can LLMs serve as practical generators of software diversity, and how much reliability improvement can that diversity actually provide? In this paper, we extend classical empirical studies of software diversity in human-written programs to LLM-generated code. We study three specifications using both historical human-written programs and large pools of LLM-generated ones evaluated under a common compilation, sandboxing, and exhaustive test suite. We explore LLM diversity along multiple axes, including model family, generation temperature, and programming language. Reliability improvement is evaluated in a 1-out-of-2 configuration across both homogeneous and heterogeneous program populations, including within-LLM pairings and pairings across programming languages and across LLM-generated and human-written programs. The results show that combining LLM-generated programs, especially in heterogeneous settings, can yield reliability gains, although this is partly conditioned by the programming language and generation setting. Taken together, these findings suggest that LLMs provide a scalable source of comparatively low-cost programs whose diversity can be leveraged for reliability improvement.

cs.SE↗

arXiv2026

Resilience under Uncertainty: Securing 6G through Stochastic Reinstantiation of RAN Functions

The disaggregation of base stations into discrete RAN functions introduces new threats to mobile networks, as failures in one RAN function can trigger cascading failures and interrupt entire function chains, with potential to degrade network performance and disrupt service. In this paper, we propose the first resilience mechanism for disaggregated mobile networks that leverages the adaptive reinstantiation of RAN functions under uncertainty to mitigate disruptions and maintain service continuity in the presence of compromised infrastructure. Our mechanism reacts to cascading failures that disrupt Radio Units (RUs) by reinstantiating Central Units (CUs) and Distributed Units (DUs) in alternative cloud locations, restoring their function chains while accounting for uncertainty in users' locations and wireless channel conditions during the in-failure state. We formulate this recovery process as a two-stage stochastic optimization problem, where reinstantiation and routing decisions are made under uncertainty, and bandwidth allocation decisions are performed after uncertainty is resolved. We solve the problem using a Sample Average Approximation (SAA)-based solution as a tractable, deterministic equivalent problem. We numerically evaluate our approach on a real-world disaggregated mobile network topology across multiple failure scenarios and traffic demand conditions, and our results demonstrate that our solution can achieve up to 80% higher recovery performance compared to conventional resilience mechanisms.

cs.NI↗

arXiv2026

Rethinking Reinforcement Fine-Tuning in LVLM: Convergence, Reward Decomposition, and Generalization

Reinforcement fine-tuning with verifiable rewards (RLVR) has emerged as a powerful paradigm for equipping large vision-language models (LVLMs) with agentic capabilities such as tool use and multi-step reasoning. Despite striking empirical successes, most notably Visual Agentic Reinforcement Fine-Tuning (Visual-ARFT), the theoretical underpinnings of this paradigm remain poorly understood. In particular, two critical questions lack rigorous answers: (i)~how does the composite structure of verifiable rewards (format compliance, answer accuracy, tool executability) affect the convergence of Group Relative Policy Optimization (GRPO), and (ii)~why does training on a small set of tool-augmented tasks transfer to out-of-distribution domains? We address these gaps by introducing the \emph{Tool-Augmented Markov Decision Process} (TA-MDP), a formal framework that models multimodal agentic decision-making with bounded-depth tool calls. Within this framework, we establish three main results. First, we prove that GRPO under composite verifiable rewards converges to a first-order stationary point at rate $O(1/\sqrt{T})$ with explicit dependence on the number of reward components and group size (\textbf{Theorem~1}). Second, we derive a \emph{Reward Decomposition Theorem} that bounds the sub-optimality gap between decomposed per-component optimization and joint optimization, providing a precise characterization of when reward decomposition is beneficial (\textbf{Theorem~2}). Third, we establish a PAC-Bayes generalization bound for tool-augmented policies that explains the strong out-of-distribution transfer observed in Visual-ARFT (\textbf{Theorem~3}).

cs.LG↗

arXiv2026

Dissipation- versus Chaos-Induced Relaxation in Non-Markovian Quantum Many-Body Systems

In interacting quantum many-body systems, relaxation toward equilibrium reflects a competition between internal chaotic dynamics and environmental dissipation. While conventional Markovian baths typically produce exponential decay, non-Markovian dissipation can give rise to more intricate behavior, including algebraic relaxation. We study an open Sachdev-Ye-Kitaev (SYK) model coupled to a pseudogapped fermionic bath, using the Keldysh formalism to compute steady-state correlations in the large-$N$ limit. Our results uncover a rich dynamical phase diagram, with regimes of bath-driven power-law relaxation, chaos-driven exponential decay, and an intermediate pre-relaxation phase where exponential decay crosses over to algebraic decay. These findings demonstrate that non-Markovian environments can qualitatively reshape relaxation mechanisms in strongly correlated quantum many-body systems.

cond-mat.stat-mech↗

arXiv2025

Toward Scalable VR-Cloud Gaming: An Attention-aware Adaptive Resource Allocation Framework for 6G Networks

Virtual Reality Cloud Gaming (VR-CG) represents a demanding class of immersive applications, requiring high bandwidth, ultra-low latency, and intelligent resource management to ensure optimal user experience. In this paper, we propose a scalable and QoE-aware multi-stage optimization framework for resource allocation in VR-CG over 6G networks. Our solution decomposes the joint resource allocation problem into three interdependent stages: (i) user association and communication resource allocation; (ii) VR-CG game engine placement with adaptive multipath routing; and (iii) attention-aware scheduling and wireless resource allocation based on motion-to-photon latency. For each stage, we design specialized heuristic algorithms that achieve near-optimal performance while significantly reducing computational time. We introduce a novel user-centric QoE model based on visual attention to virtual objects, guiding adaptive resolution and frame rate selection. A dataset-driven evaluation demonstrates that, when compared against state-of-the-art approaches, our framework improves QoE by up to 50\%, reduces communication resource usage by 75\%, and achieves up to 35\% cost savings, while maintaining an average optimality gap of 5\%. Our proposed heuristics solve large-scale scenarios in under 0.1 seconds, highlighting their potential for real-time deployment in next-generation mobile networks.

cs.NI↗

arXiv2025

Quantum Walks on Arbitrary Spatial Networks with Rydberg Atoms

Rydberg atoms provide a highly promising platform for quantum computation, leveraging their strong tunable interactions to encode and manipulate information in the electronic states of individual atoms. Key advantages of Rydberg atoms include scalability, reconfigurable connectivity, and native multi-qubit gates, making them particularly well-suited for addressing complex network problems. These problems can often be framed as graph-based tasks, which can be efficiently addressed using quantum walks. In this work, we propose a general implementation of staggered quantum walks with Rydberg atoms, with a particular focus on spatial networks. We also present an efficient algorithm for constructing the tessellations required for the staggered quantum walk. Finally, we demonstrate that our proposal achieves quadratic speedup in spatial search algorithms.

quant-ph↗

arXiv2025

Universality, Robustness, and Limits of the Eigenstate Thermalization Hypothesis in Open Quantum Systems

The eigenstate thermalization hypothesis (ETH) underpins much of our modern understanding of the thermalization of closed quantum many-body systems. Here, we investigate the statistical properties of observables in the eigenbasis of the Lindbladian operator of a Markovian open quantum system. We demonstrate the validity of a Lindbladian ETH ansatz through extensive numerical simulations of several physical models. To highlight the robustness of Lindbladian ETH, we consider what we dub the dilute-click regime of the model, in which one postselects only quantum trajectories with a finite fraction of quantum jumps. The average dynamics are generated by a non-trace-preserving Liouvillian, and we show that the Lindbladian ETH ansatz still holds in this case. On the other hand, the no-click limit is a singular point at which the Lindbladian reduces to a doubled non-Hermitian Hamiltonian and Lindbladian ETH breaks down.

cond-mat.stat-mech↗

arXiv2021

OPlaceRAN -- a Placement Orchestrator for Virtualized Next-Generation of Radio Access Network

The fifth-generation mobile evolution enables transformations on Next-Generation Radio Access Networks (NG-RAN). The RAN protocol stack is split into eight disaggregated options combined in three network units, i.e., Central, Distributed, and Radio. Besides that, further advances allow the RAN functions to be virtualized on top of general-purpose hardware, using the concept of virtualized RAN (vRAN). The Combination of NG-RAN and vRAN results in vNG-RAN, which enables the management of the disaggregated units and protocols as a set of radio functions. However, the orchestration-based placement of these radio functions is a challenging issue since the best decision can be determined by multiple constraints involving RAN disaggregation, crosshaul networks requirements, availability of computational resources, etc. This article proposes OPlaceRAN, a vNG-RAN deployment orchestrator framed within the NFV reference architecture and aligned with the Open RAN initiative. OPlaceRAN supports the dynamic placement of radio functions focusing on vNG-RAN planning and is designed to be agnostic to the placement optimization model. To validate OPlaceRAN, we developed a prototype based on up-to-date cloud-native tools to deploy RAN using containerized virtualization using the OpenAirInterface emulator and considering two distinct functional splits (options 2 and 6). The evaluation is tested as proofs-of-concept in a real computing infrastructure using two different placement solutions. Our results reveal that OPlaceRAN is an effective cloud-native solution for containerized network functions placement and agnostic to the optimization model. Additionally, OPlaceRAN is up-to-date with the most advanced vNG-RAN design and development approaches, contributing to the evolution of the fifth-generation of mobile networks.

cs.NI↗