Search for: All records

A comprehensive understanding of the topology of the electric power transmission network (EPTN) is essential for reliable and robust control of power systems. While existing research primarily relies on domain-specific methods, it lacks data-driven approaches that have proven effective in modeling the topology of complex systems. To address this gap, this paper explores the potential of data-driven methods for more accurate and adaptive solutions to uncover the true underlying topology of EPTNs. First, this paper examines Gaussian Graphical Models (GGM) to create an EPTN network graph (i.e., undirected simple graph). Second, to further refine and validate this estimated network graph, a physics-based, domain specific refinement algorithm is proposed to prune false edges and construct the corresponding electric power flow network graph (i.e., directed multi-graph). The proposed method is tested using a synchrophasor dataset collected from a two-area, four-machine power system simulated on the real-time digital simulator (RTDS) platform. Experimental results show both the network and flow graphs can be reconstructed using various operating conditions and topologies with limited failure cases. 

Requiring less data for accurate models, few-shot learning has shown robustness and generality in many application domains. However, deploying few-shot models in untrusted environments may inflict privacy concerns, e.g., attacks or adversaries that may breach the privacy of user-supplied data. This paper studies the privacy enhancement for the few-shot learning in an untrusted environment, e.g., the cloud, by establishing a novel privacy-preserved embedding space that preserves the privacy of data and maintains the accuracy of the model. We examine the impact of various image privacy methods such as blurring, pixelization, Gaussian noise, and differentially private pixelization (DP-Pix) on few-shot image classification and propose a method that learns privacy-preserved representation through the joint loss. The empirical results show how privacy-performance trade-off can be negotiated for privacy-enhanced few-shot learning. 

Federated learning (FL) is a promising strategy for performing privacy-preserving, distributed learning with a network of clients (i.e., edge devices). However, the data distribution among clients is often non-IID in nature, making efficient optimization difficult. To alleviate this issue, many FL algorithms focus on mitigating the effects of data heterogeneity across clients by introducing a variety of proximal terms, some incurring considerable compute and/or memory overheads, to restrain local updates with respect to the global model. Instead, we consider rethinking solutions to data heterogeneity in FL with a focus on local learning generality rather than proximal restriction. To this end, we first present a systematic study informed by second-order indicators to better understand algorithm effectiveness in FL. Interestingly, we find that standard regularization methods are surprisingly strong performers in mitigating data heterogeneity effects. Based on our findings, we further propose a simple and effective method, FedAlign, to overcome data heterogeneity and the pitfalls of previous methods. FedAlign achieves competitive accuracy with state-of-the-art FL methods across a variety of settings while minimizing computation and memory overhead. Code is available at https://github.com/mmendiet/FedAlign. 

Abstract Human societies are characterized by norms that restrict selfish behavior and promote cooperation. The oxytocin system is an important modulator of social behavior that may be involved in the evolution of cooperation. Oxytocin acts in both the nucleus accumbens and the anterior cingulate cortex to promote social bonding and social cohesion. Expression of theCD38andOXTRgenes is known to affect oxytocin secretion and binding, respectively, in these brain areas. The Andean highlands provide an excellent opportunity to evaluate the role of oxytocin in the evolution of cooperation. The rich archeological record spans 13,000 years of population growth and cooperative challenges through periods of highland exploration, hunting economies, agro‐pastoralism, and urbanization. Through allele trajectory modeling using both ancient and contemporary whole genomes, we find evidence for strong positive selection on theOXTRandCD38alleles linked with increased oxytocin signaling. These selection events commenced around 2.5 and 1.25 thousand years ago, placing them in the region's Upper Formative and Tiwanaku periods—a time of population growth, urbanization, and relatively low rates of violence. Along with remarkable and enduring cultural developments, increased oxytocin secretion and receptor binding in these brain areas may have facilitated large‐scale cooperation that promoted early urbanization in the Titicaca Basin of the Andean highlands.