Kim Bente

• I am interested in developing application-driven machine learning innovations, particularly for geospatial, geophysical, and Antarctic science.
• Geospatial modelling
• Physics-informed ML
• Climate, Earth Science, and Geophysics applications
• Antarctica and the Cryosphere
• Earth observation (EO)
• EO foundation models
• Active learning with Bayesian Optimisation (BO)
• Causal inference

My name is Kim Bente and I am a final year Machine Learning PhD student at the School of Computer Science, University of Sydney ,supervised by Prof Fabio Ramos (NVIDIA, University of Sydney), and A/Prof Roman Marchant (Human Technology Institute, UTS). My research interests are centred around using probabilistic machine learning and statistical computing methods to address Climate Science problems. I am particularly interested in the quantification of uncertainty to inform high-stakes decision in the climate domain. I am currently working on the application Bayesian Optimisation to Antarctic research problems, including sensor network design, data fusion and ice core drilling site determination. I am a member of the DARE (Data Analytics for Resources and Environments) ARC Training Centre, an Industrial Transformation training centre, led by the University of Sydney. I have gained extensive experience on interdisciplinary research projects, most notably collaborating with Nutrition and Dietetics research, and also working on educational data, as well as criminology data.

I have completed the Master of Data Science [with high distinction] from the University of Sydney in 2020 and I hold a Bachelor of Science from the Technical University Munich, in Management & Technology, specialising in Chemical Engineering and Finance.

Please contact me via kim.bente@sydney.edu.au

Check out my GitHub repositories for preprocessing pipelines of Antarctic datasets (e.g. Bedmap3 points, Bedmap maps, BedMachine, RACMO 2.4, GRACE-FO, MODIS MOA, NASA's MEaSUREs datasets), Physics-Informed ML models built in PyTorch, Uncertainty Quatification using GPyTorch and BOTorch for example, and much more.

• ML4Cryo (Machine Learning for the Cryosphere) is a research community initiative that connects researchers and practitioners working at the intersection of machine learning and cryospheric science. It was co-founded in 2024 by Julia Kaltenborn, Andrew McDonald, and myself.
• Visit the ML4Cryo website.
• Join the ML4Cryo community Slack channel.
• Submit to the `Machine Learning for the Cryosphere' session (CR6.1) at EGU 2026 in Vienna, Austria (3–8 May 2026).
• Submit to the `Artificial Intelligence and Machine Learning in Antarctic Science' parallel session at the 12th SCAR Open Science Conference in Oslo, Norway (8–19 August 2026).

I organise our weekly Machine Learning & Robotics reading group at the University of Sydney, led by Prof Fabio Ramos. If you're interested in joining you can sign up to the USYD ML reading group mailing list.

• Kim Bente, Roman Marchant, Fabio Ramos. "Mass Conservation on Rails - Rethinking Physics-Informed Learning of Ice Flow Vector Fields." In NeurIPS 2025 Tackling Climate Change with Machine Learning Workshop (2025). Available on arXiv.
• Contribution to Moreno-Rodenas et al. "Applications of artificial intelligence for water management." Publication via the UNESCO Intergovernmental Hydrological Programme (2025).
• Kim Bente, Roman Marchant, Fabio Ramos. "Antarctic Bed Topography Super-Resolution via Transfer Learning." In NeurIPS 2023 Tackling Climate Change with Machine Learning Workshop.
• Kim Bente, Judy Kay, Roman Marchant. "Probabilistic Machine Learning in Polar Earth and Climate Science: A Review of Applications and Opportunities." In AAAI 2022 Fall Symposium: The Role of AI in Responding to Climate Challenges (2022).
• Kim Bente*, Emma Nassif*, Alyse Davies*, Lyndal Wellard-Cole, Jisu Jung, Judy Kay, Clare Hughes et al. "The Contribution of Nutrients of Concern to the Diets of 18-to-30-Year-Old Australians from Food Prepared Outside Home Differs by Food Outlet Types: The MYMeals Cross-Sectional Study." Nutrients 14, no. 18 (2022): 3751.
  * equal contributions
• Lyndal Wellard-Cole, Alyse Davies, Juliana Chen, Jisu Jung, Kim Bente, Judy Kay, Wendy L. Watson et al. "The contribution of foods prepared outside the home to the diets of 18-to 30-year-old Australians: the MYMeals study." Nutrients 13, no. 6 (2021): 1761.

• COMING UP  —  NeurIPS 2025 Tackling Climate Change with Machine Learning workshop hosted by Climate Change AI. I will be attending NeurIPS to present my accepted workshop paper, Mass Conservation on Rails – Rethinking Physics-Informed Learning of Ice Flow Vector Fields (poster presentation, Sunday 7 December, Upper Level Room 24ABC).
  1 - 7 December 2025, San Diego, California, USA
• Polar Data Forum VI hosted by the Australian Antarctic Data Centre — I gave a talk within the Machine Learning / Artificial Intelligence theme.
  20 - 24 October 2025, Hobart, Tasmania, Australia
• EGU 2025 — I served as co-convener of the 'Machine Learning for Cryospheric Sciences' session, and also presented at session EOS4.6.
  27 April - 2 May 2025, Vienna, Austria (virtual attendance)
• Climate Informatics 2024 — I attended and presented my poster virtually at this conference.
  22 - 24 April 2024, London, UK (virtual attendance)
• EGU 2024 — I presented my poster virtually at the 'Understanding cryospheric processes in the past, present and future using data assimilation and machine learning' session.
  14 - 19 April 2024, Vienna, Austria (virtual attendance)
• NeurIPS 2023 Tackling Climate Change with Machine Learning workshop hosted by Climate Change AI. I attended NeurIPS and presented my poster.
  10 - 16 December 2023, New Orleans, Louisiana, USA
• AAAI 2022 Fall Symposium: The Role of AI in Responding to Climate Challenges hosted by Climate Change AI. I attended the symposium and gave a talk.
  17 - 19 November 2022, Arlingtion, Virginia, USA
• Information Resilience PhD School 2022 hosted by the ARC Industrial Transformation Training Centre in Information Resilience (CIRES). I presented my research as a poster and in a 3MT.
  12 - 14 September 2022, The University of Queensland (UQ), Brisbane, Queensland, Australia

• COMING UP  —  Seminar at the Scripps Polar Center, Scripps Institution of Oceanography, University of California San Diego (UCSD)
  5/12/2025 University of California San Diego, San Diego, California, USA
• Seminar at the Geology-Geophysics seminar series, School of Geosciences, University of Sydney
  5/11/2025 University of Sydney, Sydney, Australia
• Seminar at the Centre for Research in Mathematics and Data Science at Western Sydney University, hosted by Associate Professor Oliver Obst
  25/09/2025 Western Sydney University, Sydney, Australia
• Seminar at the Compute Antarctic research group at University of Tasmania, hosted by Professor Anya Reading
  13/06/2025 University of Tasmania, Hobart, Australia
• Panelist at the Information Resilience PhD School 2024 hosted by the ARC Industrial Transformation Training Centre in Information Resilience (CIRES)
  31/10/2024 University of Queensland, Brisbane, Australia
• Research talk & Panel at the Google Deepmind event organised by USYD GDSC (Google Developer Student Club) at Google Sydney
  28/06/2024 Google, Sydney, Australia
• Seminar at the National Institute for Applied Statistics Research Australia (NIASRA) at the University of Wollongong
  6/04/2024 University of Wollongong, Wollongong, Australia

• NeurIPS 2025 Tackling Climate Change with Machine Learning workshop
• Conference on Uncertainty in Artificial Intelligence (UAI) 2025
• ICLR 2025 Tackling Climate Change with Machine Learning workshop
• Conference on Uncertainty in Artificial Intelligence (UAI) 2024
• ICLR 2024 Tackling Climate Change with Machine Learning workshop

• Machine Learning PhD student (current), The University of Sydney, Australia
  'Uncertainty Quantification and Physical Priors for Geospatial Models of Antarctica', supervised by Professor Fabio Ramos and A/Prof Roman Marchant
  member of the DARE (Data Analytics for Resources and Environments) ARC Training Centre
• Master of Data Science, The University of Sydney, Australia
  completed with High Distinction in 2020
  Final research project (Capstone) 'New Perspectives on the Validation of a Mobile App for Dietary Assessment using Data Mining', supervised by Professor Judy Kay
  Faculty of Engineering Summer Research Program Scholarship 2019/2020 for the project 'Data mining of dietary patterns', supervised by Prof Judy Kay, A/Prof Irena Koprinska and A/Prof Kalina Yacef
• Graduate Certificate in Data Science, The University of Sydney, Australia
  completed with High Distinction in 2019
• Bachelor of Science in Management & Technology, Technical University of Munich, Germany
  Specialisation in Chemical Engineering
  Bachelor's Thesis about 'Knowledge Management in a Dynamic Organisation'

• Tutor for Computer Science Research Methods [INFO5993], School of Computer Science
  Semester 2 2024, postgraduate research education
• Tutor for Computer Science Research Methods [INFO5993], School of Computer Science
  Semester 1 2023, postgraduate research education
• Tutor for Human-in-the-Loop Data Analytics [DATA3406], School of Computer Science
  Semester 2 2022, undergraduate level tutoring
• Tutor for Data Analysis in the Social Sciences [DATA5207], United States Study Centre
  Intensive February 2021, postgraduate level tutoring
• Tutor for Algorithms [COMP9007], School of Computer Science
  Semester 2 2019, postgraduate level tutoring
• Course analytics for Human-in-the-Loop Data Analytics [DATA3406], School of Computer Science
  Educational data analytics, student performance feedback with SRES, development of a computer-based marking program for Semester 2 2020

• Customer Analytics and Customer Relationship Management at FOURSOURCE, a tech start-up in apparel sourcing, Berlin, Germany
• Innovation Consulting and Ideation, Junior Innovation consultant, HYVE, Munich, Germany
• Management Consulting, media sector, working student, goetzpartners advisory, Munich, Germany
• Research & Development (R&D) and International Marketing, internships, consumer goods sector, Schwarzkopf (Henkel), Hamburg, Germany
• Product development and product management, internship, consumer goods sector, Kneipp, Würzburg, Germany

• Running, Cycling, Tennis, Hiking, Swimming, Skiing
• Designing and sewing timeless garments and other textile creations
• Cooking healthy food, travelling, exploring art

The Daily Data Dose [DDD] is a compilation of my favourite data-related resources, both informative & entertaining. My shortlist includes podcasts, websites, talks and other useful things that I came across. Most of these I have recommended to a plethora of data-enthusiasts, so it was time to put a list on the web. I thank everyone who has introduced me to any of these gems. Links are included below.

• PolarGlobe 4D climate data visualisation

  
  Amazing nowcasts as well as 12 day forecasts of global wind and ocean currents for the whole globe.

• The Vanishing Gradients podcast

  
  This is my new favourite Data Science podcasts!

• The Dataframed podcast

  
  This is my previous favourite Data Science podcasts! I have listened to every of the 60 episodes at least once, soaking up insights from industry to academia. I can't recommend this enough, and if you know me I will have likely already recommended it to you.

• My compilation of Antarctic data sets

  
  A repository on my GitHub with an overview of icy data sets, frosty models and other high-latitude data collections. I focus on the Antarctic ice sheet.

• 'AI for Earth' projects by Microsoft

  
  I learnt about some of the projects from 'AI for Earth' at a very inspiring talk by Jennifer Marsman at the Microsoft reactor in Sydney. I was very fascinated by the creative use of AI, especially by the World Mosquito Program and the Whalebook. I recommend checking out the projects.

• Process Mining Coursera course by Professor Wil van der Aalst

  
  I took this course on Process Mining which is not only interesting from an algorithmic point of view but very relevant for any temporal or procedural data set. Great course!

• "Environment Variables: how software impacts climate change" - Merrin Macleod (PyCon AU 2019)

  
  This is an outstanding talk which was recommended to me by a colleague (Thanks Ben!). If you care about the environment and you work in tech, you will surely learn something new here.

• The Pudding

  
  The Pudding is website which publishes visual essays about societal and cultural topics. They are pushing the boundaries of data journalism and therefore always worth a digital visit.

• Gapminder

  
  Gapminder is a great organisation that promotes a more fact-based worldview. They also have an extensive collection of open access data sets. Furthermore, I highly recommend reading the book 'Factfulness' by Hans Rosling.

• Information is Beautiful

  
  A collection of informative and visually appealing data visualisations.

• Google dataset search tool

  
  A search engine for data sets. A very useful tool to find openly available data sets.

• Data Science by John D. Kelleher and Brendan Tierney

  
  I recommend this book to family and friends who want to understand what I do, but I also like going back to this pocket book by MIT Press to reset my own perspective for the big picture.

• Papers with code

  
  Professor Caren Han's NLP course [COMP5046] introduced me to this site, which is a great resource to stay on top of the latest ML research in different fields.

• Datasaurus, the new Anscombe's quartet

  
  An updated and more creative version of the Anscombe's quartet, including animated visualisations. This demonstrates once again that summary statistics are not enough to describe data.

• Spurious correlations blog by Tyler Vigen

  
  A good reminder that correlation ≠ causation.

• The True size of ...

  
  This is a fun web tool to understand the distortion of the size of countries, as they are displayed on the world map.

• kim.bente@sydney.edu.au
• Find me in J12 - School of Computer Science
• The University of Sydney, 2006 NSW, Australia