Aristeidis Panos

Research Associate in Machine Learning

Department of Engineering, University of Cambridge

Biography

Aristeidis (Ares) Panos is a Research Associate at the Computational and Biological Learning (CBL) Lab in the Department of Engineering at Cambridge University. He received a B.Sc. degree in computer science in 2014 (valedictorian) from Athens University of Economics and Business and then he was awarded a Ph.D. in Machine Learning in 2019 from the Department of Statistical Science at the University College London. In October 2020, he joined the Department of Statistics at the University of Warwick as a Postdoctoral Research Fellow while, at the same time, he was a member of the Trustworthy digital identity group at the Alan Turing Institute, the UK’s national institute for data science and artificial intelligence.

Interests

Probabilistic Machine Learning
Computational Statistics
Bayesian Inference
Continual Learning
Large Vision-Language Models

Education

Ph.D. in Machine Learning, 2019

University College London
B.Sc. in Computer Science, 2014

Athens University of Economics and Business

Experience

Research Associate

Department of Engineering, University of Cambridge

Feb 2022 – Present Cambridge, UK

Developing new fundamental tools for machine intelligence and machine learning.

Postdoctoral Research Fellow

Department of Statistical Science, University College London

Dec 2021 – Jan 2022 London, UK

Forecasting wind farm output using machine learning methods.

Postdoctoral Research Fellow

Department of Statistics, University of Warwick

Nov 2020 – Nov 2021 Warwick, UK

Mechanistic marked spatio-temporal point processes for large-scale data-analytic applications.

Machine Learning Consultant

GlaxoSmithKline (GSK)

Jun 2020 – Oct 2020 London, UK

Yield optimization of multi-stage chemical processes used for medicine production via machine learning techniques.

Data scientist

EasyJet

Sep 2018 – Nov 2019 Luton, UK

Implementation of a Bayesian pricing model; causal inference using machine learning algorithms; multi-echelon inventory optimization.

Internship

The Alan Turing Institute

Jun 2016 – Aug 2016 London, UK

Machine Classifiers and Similarity measures.

Ph.D. in Machine Learning

Department of Statistical Science, University College London

Sep 2015 – Oct 2019 London, UK

Extreme Multi-label Learning with Gaussian Processes.

B.Sc. in Computer Science

Department of Informatics, Athens University of Economics and Business

Sep 2009 – Jun 2014 Athens, Greece

First Class Honours (2014 Valedictorian).

Recent Publications

Filtering publications.

Aristeidis Panos (2024). Decomposable Transformer Point Processes. In the 38th Conference on Neural Information Processing Systems (NeurIPS 2024).

PDF Cite

Aristeidis Panos, Yuriko Kobe, Daniel Olmeda Reino, Rahaf Aljundi, Richard E. Turner (2023). First Session Adaptation: A Strong Replay-Free Baseline for Class-Incremental Learning. In the IEEE/CVF International Conference on Computer Vision.

PDF Cite

Aristeidis Panos, Ioannis Kosmidis, Petros Dellaportas (2023). Scalable Marked Point Processes for Exchangeable and Non-Exchangeable Event Sequences. In the 26th International Conference on Artificial Intelligence and Statistics.

PDF Cite

Constantinos Daskalakis, Petros Dellaportas, Aristeidis Panos (2022). How Good are Low-Rank Approximations in Gaussian Process Regression?. In 36th AAAI Conference on Artificial Intelligence (oral).

PDF Cite

Aristeidis Panos, Ioannis Kosmidis, Petros Dellaportas (2021). Scalable and Interpretable Marked Point Processes. In arXiv preprint.

PDF Cite

Aristeidis Panos, Petros Dellaportas, Michalis K. Titsias (2021). Large Scale Multi-Label Learning using Gaussian Processes. In Machine Learning Journal.

PDF Cite

Sebastian Bobadilla-Suarez, Christiane Ahlheim, Abhinav Mehrotra, Aristeidis Panos, Bradley C. Love (2019). Measures of Neural Similarity. In Computational Brain & Behavior Journal.

PDF Cite

Aristeidis Panos (2019). Extreme multi-label learning with Gaussian processes. In Ph.D. thesis.

PDF

Aristeidis Panos, Petros Dellaportas, Michalis K. Titsias (2018). Fully Scalable Gaussian Processes using Subspace Inducing Inputs. In arXiv preprint.

PDF Cite