Machine Learning for Marine Biodiversity Modelling

Contact person: Alexander Eiler        
Keywords: Marine Biodiversity, Machine learning, Citizen Science, image analysis, genomics
Research group: Aquatic biology and toxicology (AQUA)    
Department of Biosciences (IBV), Department of Mathematics (MI), Department of Informatics (IFI)
 

Novel technologies now provide a wealth of marine data, offering a deeper understanding of the Ocean with huge potential for marine conservation. Environmental DNA, high-throughput imaging, and acoustics unveil detailed marine biodiversity insights. Citizen Science can potentially amplify the amount of entries to global biodiversity databases. Here, we will focus on computational tools, including diverse Machine Learning procedures, to classify and predict biodiversity changes, offer  identification of individuals and species through image analysis, ecosystem classification in relation to anthropogenic impacts, correlation of biodiversity patterns with environmental factors, and forecasting changes under anthropogenic forcing like pollution and climate-induced shifts in marine food webs.

Methodological research topics:

  • Building and interpreting classifiers
  • Modeling using various ML procedures (i.e. decision trees)
  • Decision-making with ML models
  • Building a Citizen Science data acquisition and evaluation platform
  • Genomic data analysis (i.e. bioinformatics tool development)

 
Topics from natural sciences or technology:

  • Image analysis and classifier interpretation to identify marine species and individuals
  • Forecasting Biodiversity under anthropogenic forcing (i.e. marine pollution and heatwaves)
  • Reconstruct marine food web changes
  • Ecological status classification
  • Modeling fish stocks and fish quotas
  • Advising decision making in marine ecosystem conservation and restoration

Research team:

Mentoring and internship will be offered by a relevant external partner.