Theoretical physicist by training, working on machine learning, Bayesian statistics, and fundamental physics.
-
Pre-trained transformer inference for gravitational-wave time series
Foundation-style model for inference of gravitational wave signals with domain adaptation, accelerated training and performance. Based predominantly on CNN and transformer architecture.
-
GPU-based Bayesian inference for gravitational waves
A GPU/cuda-accelerated code for simulation Bayesian forecasts for gravitational-wave time series.
-
Deep learning for chaos detection
Anomaly detection model for chaotic dynamics in dynamical systems with deep convolutional networks.
https://github.com/ippocratiss/Deep-classifier-for-chaos-and-order
-
Generative-adversarial-network-for-complex-systems
Generative model which learns and reconstructs chaotic dynamical systems.
https://github.com/ippocratiss/Generative-adversarial-network-for-complex-systems
-
Deep network pipeline for Bayesian inverse modelling of neutron stars
Bayesian deep learning pipeline to solve the inverse problem of inferring the state of dense matter in neutron stars. Based predominantly on probabilistic Bayesian neural networks.
https://github.com/ippocratiss/Deep-learning-inference-of-the-neutron-star-equation-of-state
-
Weather forecasting with Graph Neural Networks
The model uses weather stations as nodes to predict the weather in the UK. It achieves > 95% accuracy on temperature forecasting.
https://github.com/ippocratiss/minRNNs-for-weather-prediction
https://github.com/ippocratiss/Weather_GNN
- Python (PyTorch, TensorFlow), Fortran
- High-performance & GPU computing (Predominantly Cuda)
- Bayesian inference and inverse modelling, time series, transformers, generative models, diffusion models, graph neural networks, LLMs, agentic AI.
A full list of publications and research statistics can be found through the following link: