Showing 1–5 of 5 results for author: Simonetti, A

Tropical methods for stable Horikawa surfaces

Authors: Jonathan David Evans, Angelica Simonetti, Giancarlo Urzúa

Abstract: There are many strata in the KSBA boundary of the moduli space of octic double planes ($K^2=2$, $p_g=3$). We use methods from tropical and toric geometry to show that only three of these correspond to surfaces with at worst quotient singularities. There are many strata in the KSBA boundary of the moduli space of octic double planes ($K^2=2$, $p_g=3$). We use methods from tropical and toric geometry to show that only three of these correspond to surfaces with at worst quotient singularities. △ Less

Submitted 4 May, 2024; originally announced May 2024.

Comments: 64 pages, 27 figures

MSC Class: 14J10; 14J17; 14J29

Deformation types of Looijenga pairs of small length

Authors: Angelica Simonetti

Abstract: Following the work already done by E. Looijenga and others, we provide an analysis of the deformation types of Looijenga (or anticanonical) pairs $(Y,D)$, where the anticanonical divisor $D$ is made of $n$ irreducible components and $6\leq n \leq9$. In doing so we also give a description of their toric models. Following the work already done by E. Looijenga and others, we provide an analysis of the deformation types of Looijenga (or anticanonical) pairs $(Y,D)$, where the anticanonical divisor $D$ is made of $n$ irreducible components and $6\leq n \leq9$. In doing so we also give a description of their toric models. △ Less

Submitted 21 November, 2023; originally announced November 2023.

Comments: Comments are welcome!

MSC Class: 14J26

Graph Neural Networks and Time Series as Directed Graphs for Quality Recognition

Authors: Angelica Simonetti, Ferdinando Zanchetta

Abstract: Graph Neural Networks (GNNs) are becoming central in the study of time series, coupled with existing algorithms as Temporal Convolutional Networks and Recurrent Neural Networks. In this paper, we see time series themselves as directed graphs, so that their topology encodes time dependencies and we start to explore the effectiveness of GNNs architectures on them. We develop two distinct Geometric D… ▽ More Graph Neural Networks (GNNs) are becoming central in the study of time series, coupled with existing algorithms as Temporal Convolutional Networks and Recurrent Neural Networks. In this paper, we see time series themselves as directed graphs, so that their topology encodes time dependencies and we start to explore the effectiveness of GNNs architectures on them. We develop two distinct Geometric Deep Learning models, a supervised classifier and an autoencoder-like model for signal reconstruction. We apply these models on a quality recognition problem. △ Less

Submitted 4 October, 2023; originally announced October 2023.

Comments: 11 pages, Comments Welcome!

Practical Statistical Considerations for the Clinical Validation of AI/ML-enabled Medical Diagnostic Devices

Authors: Feiming Chen, Hong Laura Lu, Arianna Simonetti

Abstract: Artificial Intelligence (AI) and Machine-Learning (ML) models have been increasingly used in medical products, such as medical device software. General considerations on the statistical aspects for the evaluation of AI/ML-enabled medical diagnostic devices are discussed in this paper. We also provide relevant academic references and note good practices in addressing various statistical challenges… ▽ More Artificial Intelligence (AI) and Machine-Learning (ML) models have been increasingly used in medical products, such as medical device software. General considerations on the statistical aspects for the evaluation of AI/ML-enabled medical diagnostic devices are discussed in this paper. We also provide relevant academic references and note good practices in addressing various statistical challenges in the clinical validation of AI/ML-enabled medical devices in the context of their intended use. △ Less

Submitted 2 March, 2023; originally announced March 2023.

Comments: 20 pages, 1 table

$\mathbb{Z}/2\mathbb{Z}$-Equivariant smoothings of cusp singularities

Authors: Angelica Simonetti

Abstract: Let $p\in X$ be the germ of a cusp singularity and let $ι$ be an antisymplectic involution, that is an involution such that there exists a nowhere vanishing holomorphic 2-form $Ω$ on $X\setminus \{p\}$ for which $ι^*(Ω)=-Ω$. Assume also that the involution is fixed point free on $X\setminus\{p\}$. We prove that a sufficient condition for such a singularity equipped with an antisymplectic involutio… ▽ More Let $p\in X$ be the germ of a cusp singularity and let $ι$ be an antisymplectic involution, that is an involution such that there exists a nowhere vanishing holomorphic 2-form $Ω$ on $X\setminus \{p\}$ for which $ι^*(Ω)=-Ω$. Assume also that the involution is fixed point free on $X\setminus\{p\}$. We prove that a sufficient condition for such a singularity equipped with an antisymplectic involution to be equivariantly smoothable is the existence of a Looijenga (or anticanonical) pair $(Y,D)$ that admits an involution free on $Y\setminus D$ and that reverses the orientation of $D$. This work also contains the proof of an analogue necessary and sufficient condition for the $\mathbb{Z}/2\mathbb{Z}$-equivariant smoothability of simple elliptic singularities $p\in C(E)$ with $E$ an elliptic curve of degree $d\leq 8$ and even equipped with a $\mathbb{Z}/2\mathbb{Z}$-action. △ Less

Submitted 8 January, 2022; originally announced January 2022.

Comments: Comments welcome!

MSC Class: 14B05; 14B07; 32S30