Department of Computer Science and Technology

Paul Scherer

Picture of Paul Scherer

I am a PhD student at the University of Cambridge Computer Laboratory under the supervision of Prof. Pietro Lio' and Prof. Mateja Jamnik as part of the Artificial Intelligence Group and the Computational Biology Group. I am generously funded by the W.D Armstrong Fund.

My research interests lie within the fields of information theory, machine learning, and biomedical informatics. My current research looks into developing learning algorithms applicable to irregular structured data such as graphs and domain specific applications in precision oncology. However my general interest in the design of useful inductive biases for representation learning goes beyond graph contexts. In previous years my research has focused on developing clustering algorithms on graphs, heterogenuous data integration, and data harmonization techniques.

Outside of working I enjoy motorcycling, camping, cooking, and reading basic maths.

Other Academic Projects

  • Conducting federated data analysis using DataSHIELD and R for epidemiology research (2016-2017).
  • Basic research into harmonization of heterogenuous epidemiological data (2016-2017)
  • Investigation into role privacy of multi-robot formations using generative adversarial networks (2018).
  • Currently developing an ABM-CA model of land use transformation with the Department of Land Economy at the University of Cambridge (2019).

Publications

  • Benedek Rozemberczki, Paul Scherer,Yixuan He, George Panagopoulos, Alexander Riedel, Maria Astefanoaei, Oliver Kiss, Ferenc Beres, Guzman Lopez, Nicolas Collignon, Rik Sarkar PyTorch Geometric Temporal: Spatiotemporal Signal Processing with Neural Machine Learning Models 30th ACM International Conference on Information and Knowledge Management (CIKM2021) Best Paper Award
  • Paul Scherer, Maja Trębacz, Nikola Simidjievski, Ramon Viñas, Zohreh Shams, Helena Andres Terre, Mateja Jamnik, Pietro Liò Unsupervised construction of computational graphs for gene expression data with explicit structural inductive biases OUP Bioinformatics (2021)
  • Benedek Rozemberczki, Paul Scherer, Oliver Kiss, Rik Sarkar, Tamas Ferenci Chickenpox Cases in Hungary: a Benchmark Dataset for Spatiotemporal Signal Processing with Graph Neural Networks (WWW’21: Graph Learning Benchmarks Workshop)
  • Paul Scherer, Maja Trȩbacz, Nikola Simidjievski, Zohreh Shams, Helena Andres Terre, Pietro Liò, Mateja Jamnik Incorporating network based protein complex discovery into automated model construction 15th Machine Learning in Computational Biology (MLCB20)
  • Maja Trȩbacz, Zohreh Shams, Mateja Jamnik, Paul Scherer, Nikola Simidjievski, Helena Andres Terre, Pietro Liò Using ontology embeddings for structural inductive bias in gene expression data analysis 15th Machine Learning in Computational Biology (MLCB20)
  • Paul Scherer and Pietro Lio Learning Distributed Representations of Graphs with Geo2DR ICML 2020 Workshop in Graph Representation Learning and Beyond, 2020
  • Paul Scherer, Helena Andres Terre, Pietro Lio, Mateja Jamnik Decoupling feature propagation from the design of graph auto-encoders Arxiv Preprint, 2019
  • Nikola Simidjievski, Cristian Bodnar, Ifrah Tariq, Paul Scherer, Helena Andres-Terre, Zohreh Shams, Mateja Jamnik, Pietro Liò Variational Autoencoders for Cancer Data Integration: Design Principles and Computational Practice Front. Genet. , 2019
  • S Pastorino, T Bishop, SR Crozier, C Granström, K Kordas, LK Küpers, EC O'Brien, K Polanska, KA Sauder, MH Zafarmand, RC Wilson, C Agyemang, PR Burton, C Cooper, E Corpeleijn, D Dabelea, W Hanke, HM Inskip, FM McAuliffe, SF Olsen, TG Vrijkotte, S Brage, A Kennedy, D O'Gorman, P Scherer, K Wijndaele, NJ Wareham, G Desoye, KK Ong Associations between maternal physical activity in early and late pregnancy and offspring birth size: remote federated individual level meta-analysis from eight cohort studies. BJOG, Volume 126, Issue 4, 2018.

Talks and Lectures

  • Representation learning on dynamic graphs. 2022. University of Cambridge, Spatial Analysis and Modelling Special Topics Lectures.
  • Principles of model fitting in differentiable and non-differentiable contexts. 2021. University of Cambridge, Spatial Analysis and Modelling Special Topics Lectures.
  • Outlining the importance of inductive biases in machine learning and the principles for achieving them . 2020. University of Cambridge, Spatial Analysis and Modelling Special Topics Lectures.
  • Learning Distributed Representations of Graphs and Other Discrete Structures. 2019. University of Cambridge AI Research Seminars
  • Federated Data Analysis using DataSHIELD and R. 2017. University of Cambridge MRC Epidemiology Unit, Addenbrookes Hospital
  • Weighted Clustering Algorithms for Protein-Protein Interaction Networks 2016, University of Edinburgh

Teaching and Supervisions

  • Supervisor 1B Artificial Intelligence Easter 2019
  • Demonstrator 1A Scientific Computing Lent 2020
  • Supervisor MPhil RM03 Spatial Analysis and Modelling Lent 2020
  • Supervisor MPhil RM03 Spatial Analysis and Modelling Lent 2021
  • Supervisor/Lecturer MPhil RM03 Spatial Analysis and Modelling Lent 2022
  • Supervisor/Assistant L45 Graph Representation Learning Lent 2022

For MPhil/Part 3 2021-2022 Students

I will be co-supervising (with Prof Lio) a couple of projects on graph representation learning and biomedical data integration. Email directly to have a chat.

Contact

Paul Scherer
Department of Computer Science and Technology
University of Cambridge
15 JJ Thomson Avenue
Cambridge CB3 0FD

Last updated 2021/03/09