Merge branch 'main' of https://github.com/UW-Madison-DataScience/ML-X-Nexus

Author: qualiaMachine

Applications/Videos/Forums/mlx_2023-10-10.qmd

@@ -16,6 +16,7 @@ categories:
   - UW-Madison
   - Time-series
   - Genomics
+  - Biology
   - Healthcare
 ---
 

Applications/Videos/Other/CrossLabsAI-CrossRoads45-METL-Biophysics-based-Protein-Language-Model.qmd

@@ -15,6 +15,7 @@ categories:
   - Cross Labs AI
   - UW-Madison
   - Transfer learning
+  - Biology
   - Biophysics
   - Protein language models
   - Foundation models

Toolbox/Data/iNaturalist.qmd

@@ -46,11 +46,11 @@ iNat images have powered major ML benchmarks and tools, including the [INQUIRE](
 
 iNat data is accessible through the [iNaturalist API](https://api.inaturalist.org/), the [Global Biodiversity Information Facility (GBIF)](https://www.gbif.org/), and various competition archives. Data licensing follows Creative Commons guidelines, and attribution to individual observers is required.
 
+## Questions?
+
+If you're using iNat or want to ask about its structure or use cases, feel free to post in the [ML+X Nexus Q&A forum](https://github.com/UW-Madison-DataScience/ML-X-Nexus/discussions/categories/q-a). 
+
 ## See also
 
 - [**Data:** INQUIRE Benchmark](https://uw-madison-datascience.github.io/ML-X-Nexus/Toolbox/Data/INQUIRE.html): Built on iNat24, this benchmark supports multimodal ecological retrieval tasks.
 - [**Talk:** Automating Scientific Discovery](https://uw-madison-datascience.github.io/ML-X-Nexus/Applications/Videos/Forums/mlx_2025-03-04.html): Learn how iNat data is being used to train AI systems that support ecological research.
-
-## Questions?
-
-If you're using iNat or want to ask about its structure or use cases, feel free to post in the [ML+X Nexus Q&A forum](https://github.com/UW-Madison-DataScience/ML-X-Nexus/discussions/categories/q-a). 

Toolbox/Models/DeepForest.qmd

@@ -4,7 +4,7 @@ author:
   - name: Chris Endemann
     email: [email protected]
 
-date: 2025-04-03
+date: 2025-04-04
 date-format: long
 image: "../../../images/DeepForest.png"
 
@@ -49,7 +49,7 @@ DeepForest builds on progress in object detection and remote sensing for ecologi
 ## Model playground
 
 All pretrained models are currently based on RetinaNet and convolutional neural networks (CNNs); no ViTs or transformer-based models (yet).
-Pretrained  models are available via Hugging Face:
+Pretrained  models are available via Hugging Face & GitHub: [huggingface.co/weecology](https://huggingface.co/weecology)
 
 - `weecology/deepforest-tree`: General tree crown detection  
 - `weecology/deepforest-bird`: Aerial bird detection  
@@ -80,9 +80,9 @@ This makes it easy to adapt DeepForest to new regions, forest types, or species
 
 ## Performance expectations
 
-- Tree crown model: F1-scores between 0.73 and 0.95 depending on site and canopy structure ([Weinstein et al. (2019), Remote Sensing, DOI: 10.3390/rs11111309](https://www.mdpi.com/2072-4292/11/11/1309))
-- Bird model: ~65% recall on new data; ~84% recall with just 1,000 local annotations  ([Weinstein et al. (2022), Ecological Applications, DOI: 10.1002/eap.2694](https://doi.org/10.1002/eap.2694))
-- Alive/dead classifier: 95.8% accuracy on held-out image crops ([Hugginf Face model card](https://huggingface.co/weecology/cropmodel-deadtrees)) 
+- Tree crown model: F1-scores between 0.73 and 0.95 depending on site and canopy structure [[Weinstein et al. (2019), Remote Sensing, DOI: 10.3390/rs11111309](https://www.mdpi.com/2072-4292/11/11/1309)]
+- Bird model: ~65% recall on new data; ~84% recall with just 1,000 local annotations  [[Weinstein et al. (2022), Ecological Applications, DOI: 10.1002/eap.2694](https://doi.org/10.1002/eap.2694)]
+- Alive/dead classifier: 95.8% accuracy on held-out image crops [[Hugging Face model card](https://huggingface.co/weecology/cropmodel-deadtrees)]
 
 Performance varies by imagery quality, tree species, and geographic region. Fine-tuning is recommended for most new applications.