Past events

2024

26.09.2024 at 14:15 CoE: Virtual laboratory for molecular level atmospheric transformations (VILMA) - Sub-seminar.

• 14.15–15.15 Joonas Vanhanen, Airmodus Ltd.: Measuring new particle formation using condensation particle counting technique
• 15.15–15.30 Coffee break

30.05.2024 at 14:15 CoE: Virtual laboratory for molecular level atmospheric transformations (VILMA) - Sub-seminar.

• 14.15–15.15 Prof. Henrik Kjærgaard, University of Copenhagen: Hydrogen Bonds and Autoxidation is everywhere in the Atmosphere 
• 15.15–15.30 Coffee break
• 15.30–16.00 Hilda Sandström, Aalto University: Molecular descriptor engineering for machine learning predictions in atmospheric science

04.04.2024 at 14:15 CoE: Virtual laboratory for molecular level atmospheric transformations (VILMA) - Sub-seminar.

• 14.15–15.15 H. J. Jost, Karsa Oy: “Multi-scheme chemical ionization mass spectrometry for atmospheric applications and beyond.”
• 15.15–15.30 Coffee break
• 15.30–16.00 Nina Sarnela, University of Helsinki: ”Towards more standardized measurements of condensable vapours.”

07.03.2024 at 14:15 CoE: Virtual laboratory for molecular level atmospheric transformations (VILMA) - Sub-seminar.

• 14.15–15.15 Coty Jen, Carnegie Mellon University: "Nucleation potential to model and measure nucleation precursors."
• 15.15–15.30 Coffee

08.02.2024 at 14:15 CoE: Virtual laboratory for molecular level atmospheric transformations (VILMA) - Sub-seminar.

• 14.15–15.15 Dr. Jarmo Mäkelä, CSC – IT Center for Science: “Digital twins and how they differ from traditional modelling.”
• 15.15–15.30 Coffee break
• 15.30–16.00 Lauri Franzon, University of Helsinki: "When does Peroxy Radical Recombination lead to low-volatility products? New insights from a qualitative reaction pathway analysis."

2023

02.02.2023 at 14:15 CoE: Virtual laboratory for molecular level atmospheric transformations (VILMA) - Sub-seminar 1.

• Siddharth Iyer, Tampere University: "Molecular rearrangement of bicyclic peroxy radicals: key route to aerosol from aromatics."
• Vitus Besel, University of Helsinki: "Curation of Big Data for Atmospheric Science."

16.03.2023 at 14:15 CoE: Virtual laboratory for molecular level atmospheric transformations (VILMA) - Sub-seminar 2. 

• 14.15-15.15 Otso Ovaskainen, University of Jyväskylä: "Surveying the world’s biodiversity with DNA, audio, image, machine learning, statistics and mathematics." (Via Zoom)
• 15.15-15.30 Coffee break
• 15.30-16.00 Timo Pekkanen, University of Helsinki: "The Chemical Master Equation". (On-site)

04.05.2023 at 14:15 CoE: Virtual laboratory for molecular level atmospheric transformations (VILMA) - Sub-seminar 3.

• 14.15-15.15 Jari Kaipio, University of Eastern Finland: "What makes a problem an unstable (inverse) one and how to interpret it?" (On-site)
• 15.15-15.30 Coffee break
• 15.30-16.00 Aku Seppänen, University of Eastern Finland: "Tomographic imaging of greenhouse gases using open-path laser dispersion spectroscopy." (Tentatively via Zoom)

01.06.2023 at 14:15 CoE: Virtual laboratory for molecular level atmospheric transformations (VILMA) - Sub-seminar 4. 

• 14.15-15.15 Juho Rousu, Aalto University: "Machine Learning for small molecule identification from mass spectrometry data."
• 15.15-15.30 Coffee break
• 15.30-16.00 Ivo Neefjes, University of Helsinki: "Configurational sampling for ion mobility modeling."

21.09.2023 at 14:15 CoE: Virtual laboratory for molecular level atmospheric transformations (VILMA) - Sub-seminar. 

• 14.15-15.15 Thomas Berkemeier, Max Planck Institute for Chemistry: "Multiphase Kinetics Models and Machine Learning in Aerosol Chemistry."
• 15.15-15.30 Coffee break
• 15.30-16.00 Bernhard Reischl, Computational Aerosol Physics group, INAR/Physics, University of Helsinki: "Unraveling the details of heterogeneous ice nucleation on silver iodide surfaces with atomistic simulations."

12.10.2023 at 14:15 CoE: Virtual laboratory for molecular level atmospheric transformations (VILMA) - Sub-seminar.

• 14.15-15.15 Prof. Claudia Mohr, Paul Scherrer Institute: "Aerosol chemical composition in the context of climate change and air quality mitigation."
• 15.15-15.30 Coffee break
• 15.30-16.00 Henning Finkenzeller, University of Helsinki: “Towards the ideal CIMS inlet – physical processes and limitations.”

16.11.2023 at 14:15 CoE: Virtual laboratory for molecular level atmospheric transformations (VILMA) - Sub-seminar.

• 14.15-15.15 Assistant Professor Jonas Elm, Aarhus University: "Machine Learning Approaches for Studying Atmospheric Molecular Cluster Formation."
• 15.15-15.30 Coffee break
• 15.30-16.00 Christopher Daub, University of Helsinki: "How do external electric fields affect the structure and stability of atmospherically relevant clusters? A DFT study."

Date and time: Thursday, 11 April 2024 at 09:00-12:15

Venue: Helsinki Expo and Convention Centre (Messuaukio 1, 00520 Helsinki)

What We Know (and Don’t Know) about Climate Change

Chair: Arkke Eskola

9:30–9:40 Session Opening

Arkke Eskola, Senior University Researcher, University of Helsinki

9:40–10:15 Soil Carbon and Climate Change

Kristiina Karhu, Assistant Professor in Soils and Climate Change, University of Helsinki

10:15–10:50 How Atmospheric Chemistry Forms Extremely Low-Volatile Organic Vapors

Mikael Ehn, Professor, University of Helsinki

10:50–11:25 Climate Change – Knowns and Unknowns

Joonas Merikanto, Senior Researcher, Finnish Meteorological Institute

11:25–11:55 The Hidden Role of Gases in Trees

Stephen Ingram, PostDoctoral Researcher, University of Helsinki

Turin, Italy, 18 September 2023 

https://sites.google.com/view/maclean23/

The workshop is part of ECML PKDD 2023 

KEY DATES

12 June 2023: Paper submission.

12 July 2023: Notification of acceptance.

CONTEXT

The vast amount of data currently produced by modern Earth Observation (EO) missions and measurements on the surface has raised new challenges for the Remote Sensing Community and atmospheric modellers. EO sensors can now offer (very) high spatial resolution images with revisit time frequencies never achieved before considering different signals, e.g., multi(hyper)spectral optical, radar, LiDAR, and Digital Surface Models. 

On the other hand, atmospheric composition and processes are measured on the surface, starting from molecular scale measurements with mass spectrometers, particle counters, and more traditional meteorological instruments. Modern machine learning techniques can be crucial in dealing with such heterogeneous, multi-scale, and multi-modal data. 

Some methods gaining attention in this domain include deep learning, domain adaptation, semi-supervised approach, time series analysis, active learning, explainable artificial intelligence, uncertainty quantification, and interactive model building and visualisation. Even though machine learning and the development of ad-hoc techniques are gaining popularity, we still see a significant need for more interaction between domain experts and machine learning researchers. 

This workshop aims to be an international forum where machine learning researchers and domain experts can meet each other to exchange, debate, and draw short and long-term research objectives around the exploitation and analysis of EO and atmospheric data via Machine Learning techniques. Among the workshop’s goals, we want to give an overview of the current machine-learning research dealing with EO and other atmospheric measurement data. On the other hand, we want to stimulate concrete discussions to pave the way to new machine learning frameworks especially tailored to deal with such data.

TOPICS

The non-exclusive list of topics for the workshop includes, to the extent related to the EO and atmospheric processes: 

• Supervised and unsupervised machine learning methods 
• Semi-supervised classification, domain adaptation, active learning, structured output learning, multi-task learning, and online learning
• Interpretability and explainability of machine learning methods
• Bayesian modelling of various parts of EO or atmospheric processes
• Dimensionality reduction and feature selection, finding embeddings and latent variables
• Visualisation and interaction with EO and atmospheric data
• Interactive model building and eliciting expert knowledge
• Applications of high-performance computing 

SUBMISSION AND MORE INFORMATION

We welcome original contributions, either theoretical or empirical, describing ongoing

projects or completed work. Contributions can be of two types: short position papers (up to 6 pages, including references) or full research papers (up to 10 pages, including references). Papers must be written in LNCS format, i.e., accordingly to the ECML-PKDD 2023 submission format. Accepted contributions will be made available electronically through the Workshop web page. Springer will publish the post-proceedings.

See the workshop website at https://sites.google.com/view/maclean23/ for more information and updates. See the ECML PKDD 2023 website at https://2023.ecmpkdd.org/ for more details about the venue.

PROGRAM COMMITTEE CHAIRS

Thomas Corpetti, CNRS, LETG-Rennes COSTEL UMR 6554 CNRS, Rennes, France

Dino Ienco, INRAE, UMR Tetis, Montpellier, France

Roberto Interdonato, CIRAD, UMR Tetis, Montpellier, France

Minh-Tan Pham, Univ. Bretagne-Sud, UMR 6074, IRISA, Vannes, France

Patrick Rinke, Aalto University, Helsinki

Kai Puolamäki, University of Helsinki, Helsinki, Finland

Monday, January 16, 2023

14:00-15:00

Place of Seminar: Kumpula, Exactum D122 (in person) & Zoom (Meeting ID: 640 5738 7231 ; Passcode: 825217)

Associate Professor Kai Puolamäki will give a talk in the Machine Learning Coffee Seminar of the Finnish Center for Artificial Intelligence (FCAI).

Abstract: I will provide an overview of our new Virtual Laboratory for Molecular Level Atmospheric Transformations (VILMA) Centre of Excellence and discuss why explainable AI (XAI) and quantifying uncertainties is essential for us. As an example of our ongoing work, I will describe SLISEMAP, a supervised manifold visualisation method, a technique for XAI developed by us that finds local explanations for all data items. SLISEMAP produces (typically) two-dimensional global visualisation of the black box model such that data items with similar local explanations will be embedded nearby.