Publications

SIGNN - Star Identification using Graph Neural Networks

Neural Network

Astrophysics

Computer Vision

Research

Conference

As a solution for the lost-in-space star identification problem we present Star Identification using Graph Neural Network (SIGNN), a novel approach using Graph Attention Networks.

Floyd Hepburn-Dickins, Mark W. Jones, Mike Edwards, Jay Paul Morgan, Steve Bell

A first vocal repertoire characterization of long-finned pilot whales (Globicephala melas) in the Mediterranean Sea: a machine learning approach

Research

Neural Network

Bioacoustics

Journal

The acoustic repertoires of long-finned pilot whales (Globicephala melas) in the Mediterranean Sea are poorly understood. This study aims to create a catalogue of calls, analyse acoustic parameters, and propose a classification tree for future research.

Marion Poupard, Paul Best, Jay Paul Morgan, Gianni Pavan, Hervé Glotin

Removing cloud shadows from ground-based solar imagery

Research

Neural Network

Heliophysics

Image Processing

Computer Vision

Astrophysics

Journal

Cloud coverage can contaminate and occlude our vision of the Sun. This cloud-contamination can vastly reduce the amount of usable data for scientists studying the weather conditions on the Sun. In this work, we address this issue using Neural Networks to remove clouds from solar imagery.

Amal Chaoui, Jay Paul Morgan, Adeline Paiement, Jean Aboudarham

Domain-informed graph neural networks: a quantum chemistry case study

Research

Quantum Chemistry

Neural Network

Journal

Estimating the potential energies of out-of-equilibrium molecules and crystals using expert-informed Neural Networks.

Jay Paul Morgan, Adeline Paiement, Christian Klinke

Removing cloud shadows from ground-based solar imagery

Research

Heliophysics

Neural Network

Preprint

A U-Net-style Neural Network to remove cloud contaminants from ground-based solar imagery.

Amal Chaoui, Jay Paul Morgan, Adeline Paiement, Jean Aboudarham

Strategies to use prior knowledge to improve the performance of Deep Learning: an approach towards trustable Machine Learning systems

Thesis

PhD Thesis of developing strategies for incorporating priors and domain knowledge into design of Deep Neural Networks.

Jay Paul Morgan

A Computability Perspective on (Verified) Machine Learning

Conference

Computability

Adversarial Examples

Define the computational tasks underlying the newly suggested verified ML in a model-agnostic way, i.e., they work for all machine learning approaches including, e.g., random forests, support vector machines, and Neural Networks.

Tonicha Crook, Jay Paul Morgan, Arno Pauly, Markus Roggenbach

Adaptive Neighbourhoods for the Discovery of Adversarial Examples

Preprint

Adversarial Examples

Constructing unique and informed bounds for constructing adversarial examples.

Jay Paul Morgan, Adeline Paiement, Arno Pauly, Monika Seisenberger

A Chatbot Framework for the Children’s Legal Centre

Conference

Neural Network

This paper presents a novel method to address legal rights for children through a chatbot framework by integrating machine learning, a dialogue graph, and information extraction.

Jay Paul Morgan, Adeline Paiement, Monika Seisenberger, Jane Williams, Adam Wyner