Bayesian palaeoclimate reconstruction from zero-inflated count-compositional pollen data: A case study of Lago Grande di Monticchio in southern Italy
本文提出了一种贝叶斯模块化框架，通过零膨胀计数-组合花粉数据重建古气候，解决了地理大范围校准数据集引入的异质性和复杂花粉-气候关系问题。
0 citationsRead paper
Selected work
Representative Papers
Performance Evaluation of HAPS-enabled Coverage Enhancement in Hard-to-Reach Areas
该论文使用随机几何工具建模难以到达的区域，并通过部署高空平台站（HAPS）来增强这些地区的网络覆盖，分析了不同参数对上下行链路覆盖性能的影响。
0 citationsRead paper
Connectivity of HAPS-Based Solutions for Large-Scale Wireless Networks: A Percolation Theory Analysis
本文使用渗流理论分析了基于高空平台站（HAPS）的三种方案，研究其在无光纤区域实现大规模连续互联网覆盖的可行性。
0 citationsRead paper
Tabular foundation models for non-tabular tasks
研究使用TabPFN v3模型解决非表格数据分类问题，将图像和文本数据转化为表格形式进行处理，无需额外训练即可达到与专用模型相近的准确率。
0 citationsRead paper
EBGT: Epistemology-aided Bayesian Game Theory for Uplink Power Control in Stochastically Distributed IoT Tiers
This work addresses the challenges of power control in dense heterogeneous IoT uplink scenarios, where conventional methods struggle due to incomplete channel state information, inter-device interference, and stringent size, weight, and power (SWaP) constraints. To overcome these limitations, the authors propose a novel epistemic Bayesian game-theoretic framework that, for the first time, incorporates cognitive hierarchy modeling into power control. The approach employs a two-layer belief structure to capture users’ reasoning about opponents’ strategies and their own strategic self-assessment. Spatial randomness is modeled using Poisson point processes and stochastic geometry, while decentralized power optimization is achieved without feedback by leveraging Jensen–Shannon divergence analysis. Theoretical derivations yield a coverage probability-based utility function, and simulations demonstrate that, under high network density and stringent SINR requirements, the proposed method significantly reduces transmit power compared to fractional power control (FPC) and signal-to-noise-and-channel-power control (SNCPC), while maintaining the target coverage probability.
0 citationsRead paper
Recent publications
Latest Papers
Bayesian palaeoclimate reconstruction from zero-inflated count-compositional pollen data: A case study of Lago Grande di Monticchio in southern Italy
本文提出了一种贝叶斯模块化框架，通过零膨胀计数-组合花粉数据重建古气候，解决了地理大范围校准数据集引入的异质性和复杂花粉-气候关系问题。
0 citationsRead paper
Performance Evaluation of HAPS-enabled Coverage Enhancement in Hard-to-Reach Areas
该论文使用随机几何工具建模难以到达的区域，并通过部署高空平台站（HAPS）来增强这些地区的网络覆盖，分析了不同参数对上下行链路覆盖性能的影响。
0 citationsRead paper
Connectivity of HAPS-Based Solutions for Large-Scale Wireless Networks: A Percolation Theory Analysis
本文使用渗流理论分析了基于高空平台站（HAPS）的三种方案，研究其在无光纤区域实现大规模连续互联网覆盖的可行性。
0 citationsRead paper
Tabular foundation models for non-tabular tasks
研究使用TabPFN v3模型解决非表格数据分类问题，将图像和文本数据转化为表格形式进行处理，无需额外训练即可达到与专用模型相近的准确率。
0 citationsRead paper
EBGT: Epistemology-aided Bayesian Game Theory for Uplink Power Control in Stochastically Distributed IoT Tiers
This work addresses the challenges of power control in dense heterogeneous IoT uplink scenarios, where conventional methods struggle due to incomplete channel state information, inter-device interference, and stringent size, weight, and power (SWaP) constraints. To overcome these limitations, the authors propose a novel epistemic Bayesian game-theoretic framework that, for the first time, incorporates cognitive hierarchy modeling into power control. The approach employs a two-layer belief structure to capture users’ reasoning about opponents’ strategies and their own strategic self-assessment. Spatial randomness is modeled using Poisson point processes and stochastic geometry, while decentralized power optimization is achieved without feedback by leveraging Jensen–Shannon divergence analysis. Theoretical derivations yield a coverage probability-based utility function, and simulations demonstrate that, under high network density and stringent SINR requirements, the proposed method significantly reduces transmit power compared to fractional power control (FPC) and signal-to-noise-and-channel-power control (SNCPC), while maintaining the target coverage probability.
0 citationsRead paper