6 papers
Critical Challenges and Guidelines in Evaluating Synthetic Tabular Data: A Systematic Review
Nazia Nafis, Inaki Esnaola, Alvaro Martinez-Perez +2
Generating synthetic tabular health data is challenging, and evaluating their quality is equally, if not more, complex. This systematic review highlights the critical importance of…
Decentralized Machine Learning with Centralized Performance Guarantees via Gibbs Algorithms
Yaiza Bermudez, Samir M. Perlaza, Iñaki Esnaola
In this paper, it is shown, for the first time, that centralized performance is achievable in decentralized learning without sharing the local datasets. Specifically, when clients…
Symmetry of concentration and scaling for self-bounding functions
George Crowley, Iñaki Esnaola
We prove generalised concentration inequalities for a class of scaled self-bounding functions of independent random variables, referred to as self-bounding. The scaling…
Solar Vortices as Conduits for Magnetoacoustic Waves: Multi-Layer Coupling and Their Role in Atmospheric Heating
Suzana S. A. Silva, Ioannis Dakanalis, Luiz A. C. A. Schiavo +10
The Sun's atmosphere hosts swirling plasma structures, known as solar vortices, which have long been thought to channel wave energy into higher layers. Until now, no direct observa…
TaylorPODA: A Taylor Expansion-Based Method to Improve Post-Hoc Attributions for Opaque Models
Yuchi Tang, Iñaki Esnaola, George Panoutsos
Post-hoc model-agnostic local attribution (LA) methods have been widely adopted to explain opaque AI models by quantifying feature-wise contributions. However, many existing method…
Submodularity of Mutual Information for Multivariate Gaussian Sources with Additive Noise
George Crowley, Inaki Esnaola
Sensor placement approaches in networks often involve using information-theoretic measures such as entropy and mutual information. We prove that mutual information abides by submod…