3ISMLG is the third International Symposium on Machine Learning and Big Data in Geosciences since its beginning in 2018. This years' conference will be conducted in a hybrid mode, meaning it is possible to attend both in Wroclaw, Poland, and for free online.

The conference is organised by the the ISSMGE Technical Committee 309: ''Machine Learning and Big Data in Geotechnics'', where Zhongqiang Liu from NGI is Chairman. Tom F. Hansen from NGI will give a keynote lecture and is responsible for the machine learning competition.

- We hope that as many as possible will attend to see the latest advancements within machine learning in geoengineering, says Zhongqiang and Tom.

View the conference program here.

Machine learning competition

There has also been a machine learning competition going on for a few months, organised on the Kaggle platform. The competition is now closed. The top 3 competition teams are invited to the event to present their solutions in a special event. Based on the Kaggle score, an extended abstract and the presentation, a winner will be announced during the conference.

Tom has led the competition and organised it together with international partners. Along with developers at NGI Digital, he has curated a hydrogeological dataset from the planned railway and road project, Ringeriksbanen. The goal of the competition was to forecast pore-pressure- and water-level sensor measurements.

Geotechnical engineers (students and practitioners) were all invited to take part in the competition.

From workshop to conference

Today, the conference is the biggest of its kind within Machine Learning (ML) in geoengineering. And it is still growing.

The first conference was a workshop with 40 participants held at NGI in Norway in 2018. The second conference was held in Shanghai, China, with around 170 participants in 2019.

- There has been a steady growth in participants since we started in 2018. We are excited to see how many attendees there will be this year. Our expectation is a few hundreds, especially since the online event is free of charge, says Zhongqiang.

Machine learning at NGI

There are many people across sections working in Machine learning related projects at NGI. It has also been established a technical group focusing on the use of machine learning in geotechnical applications.

Some examples of projects are: forecasting landslide displacements, predicting landslide spatial distribution, segmenting snow avalanche run-out areas, predicting CO2 leak-off pressure in offshore wells, interpreting CPT-measurements, rockmass characterization in conventional tunnelling from Measure While Drilling sensor-data.