Avatar

Kenneth Thorø Martinsen

Biologist (PhD)

About

Freshwater biologist doing research on machine learning and large scale carbon cycling. Highly interested in everything related to data, statistics, geospatial analysis, and machine learning. I have developed software, methods, and equipment to quantify and analyze environmental conditions. Currently, I explore how the combination of large data sets and machine learning can improve our understanding of land-water ecosystem interactions.

This webpage is an outlet for posts about ongoing and previous projects. Everything is likely related to data and water in some way!

Interests

  • Data and machine learning
  • GIS
  • Carbon cycling
  • R, Python, Linux and open-source software

Education

  • PhD in Biology, 2021

    University of Copenhagen

  • MSc in Biology, 2016

    University of Copenhagen

  • BSc in Biology, 2014

    University of Copenhagen

Recent Posts

Large language models for text translation

In recent years, machine translation has come a long way. Thanks to advances in artificial intelligence and natural language processing …

Finetuning GPT-2 for scientific text generation

Suggesting that deep learning models based are capable of generating realistic text from a prompt would be an understatement. Ever …

Deploy machine learning models with R Shiny and ONNX

Python is often the go-to language for machine learning, especially for training deep learning models using the PyTorch or TensorFlow …

Plant ID app (part 2): REST API

In part 1 of this blog post, we downloaded ~25.000 images of 100 plant species and trained a deep learning classification model. The …

Plant ID app (part 1): Data and model training

Plants species can be truly difficult to tell apart and this job often requires expert knowledge. However, when images are available …

Projects

Predicting lake bathymetry using deep learning

Estimating depths in lakes using machine learning

National database of lake catchments

Delineating lake catchments for all (>180 k) Danish lakes

Sensing lakes and streams using machine learning

Predicting lake and stream chemistry from catchment characteristics using machine learning

Open-source software for analyzing sonar data

Developing the ‘sonaR’ R-package for reading and processing sonar data

Predicting carbon dioxide in Scandinavian stream networks

Using machine learning and GIS to predict carbon dioxide in streams

Fish death in Lake Filsø

Investigating the causes of massive sudden fish death in Lake Filsø, Denmark.

New method for measuring carbon dioxide flux

Developing an automatic floating chamber for carbon dioxide flux measurements

Towards a national freshwater carbon budget

Quantifying large scale carbon dioxide concentration and emission from freshwaters

Contact