You need to read the paper plus submit a small written reflection. The reflection should be pushed to github before class. You won't get credit for the discussion unless you push to github before class. The reflection should include your thoughts on each of the questions under "questions to consider" below. It can be very brief and in bullet-point form but feel free to expand if you prefer.
Monday
- Christin S, Hervet É, Lecomte N (2019). Applications for deep learning in ecology. Methods in Ecology and Evolution 10:1632–1644. https://doi.org/10.1111/2041-210X.13256.
- This is a review of deep learning applications to provide an overview
- Questions to consider:
- What applications of deep learning strike you as likely to make the greatest impact in ecology and why?
- What applications are particularly important for your own research and how could you apply them?
- (Optional) Can you think of any applications not covered by this review?
- 2-3 other bullet points, which might be insights or questions you want to raise in the discussion
Wednesday
- Valavi R, Guillera-Arroita G, Lahoz-Monfort JJ, Elith J (2021). Predictive performance of presence-only species distribution models: a benchmark study with reproducible code. Ecological Monographs 0:e01486. https://doi.org/10.1002/ecm.1486.
- This is a case study using a range of machine learning algorithms, including random forest and boosting (but not deep learning)
- Questions to consider:
- What modeling approaches were found to be best?
- What is it about those algorithms that likely made them the best performers?
- Can the performance of any of the algorithms be improved (and likely by how much)?
- What could be the impact of sample size on the conclusions from this paper?
- Do you have any critiques of the paper, either methodological or about the conclusions drawn?
- 2-3 other bullet points, which might be insights or questions you want to raise in the discussion