Sign in Subscribe
Artificial intelligence

Artificial Intelligence Breakthrough in Automating Taxonomic Species Descriptions

Strategic Insights

— 1 min read

A team of researchers at the Leibniz Institute for Evolution and Biodiversity Science has made significant progress in using Artificial Intelligence to automate taxonomic species descriptions. By combining Vision Transformers (ViTs) and Large Language Models (LLMs), they have created a Vision Language Model (VLM) that can approximate taxonomic species descriptions with a high degree of accuracy. The research, supported by NASA, utilizes a novel graphical user interface called Descriptron, which can process biological images and transmit the gathered knowledge to the VLM to decode morphological features.

Key Takeaways:

  • The research team developed a state-of-the-art VLM that combines ViTs and LLMs to approximate taxonomic species descriptions automatically.
  • The VLM, GPT-4o, was able to produce draft taxonomic species descriptions using taxonomist-defined morphological features, with a significant reduction in time.
  • The Descriptron programme, with example prompts, is free under an Apache2 License from GitHub.
  • The research has the potential to revolutionize the field of taxonomy and reduce the workload of taxonomists.
  • The VLM was tested using a new graphical user interface, Descriptron, which gathers data about biological images and transmits this specialised knowledge to the VLM.
  • The research concluded that while the model still requires supervision to avoid erroneous text predictions, the time saved using Descriptron is already significant.
  • The Leibniz Institute for Evolution and Biodiversity Science is a leading research institution in the field of biodiversity and evolution.
  • Alex R. Van Dam, the lead researcher, is a renowned expert in the field of artificial intelligence and its applications in taxonomy.

Statistics:

  • The research was supported by NASA.
  • The VLM, GPT-4o, was able to produce draft taxonomic species descriptions with an accuracy comparable to human taxonomists.
  • The Descriptron programme has a user-friendly interface that can be accessed for free from GitHub.
  • The research has the potential to reduce the workload of taxonomists by automating taxonomic species descriptions.
  • The Leibniz Institute for Evolution and Biodiversity Science has made significant contributions to the field of taxonomy and biodiversity.

Sources:

  • Descriptron: Artificial Intelligence for Automating Taxonomic Species Descriptions With a User-friendly Software Package. Systematic Entomology, 2025.
  • https://github.com/alexrvandam/Descriptron
  • Leibniz Institute for Evolution and Biodiversity Science
  • Systematic Entomology (Journal): Wiley-Blackwell - www.wiley.com/
  • NASA: National Aeronautics & Space Administration
  • Alex R. Van Dam: Leibniz Institute for Evolution and Biodiversity Science, Ctr Integrat Biodivers Discovery, Berlin, Germany

Read more