Add foundation models page (#354)

* Add foundation models page

* Update CHANGELOG

* Update CHANGELOG.md

---------

Co-authored-by: Robrecht Cannoodt <[email protected]>

CHANGELOG.md

@@ -12,6 +12,8 @@
 
 * Update base images documentation (PR #342).
 
+* Add foundation models benchmark page (PR #354)
+
 ## BUG FIXES
 
 * Fix error in summary figure when not all method info ids are present in the results (PR #330).
@@ -34,12 +36,12 @@
 
 ## NEW FUNCTIONALITY
 
-* Add gloassary function to pages and a glossary page (PR #300). 
+* Add glossary function to pages and a glossary page (PR #300).
 
 ## MAJOR CHANGES
 
 * Update `fundamentals` pages (PR #300).
-  
+
 * Refactor the reference view (PR #300).
 
 * Update `reference` documentation (PR #300).
@@ -85,7 +87,7 @@
 * Update neurips 2021 datasets data with new dataset_id (PR #315).
 
 * Update to funkyheatmapjs v0.2.5 (PR #316).
-  
+
 * Add CSS to fix not active tab color in dark mode (PR #319).
 
 * Improvements to the documentation (PR #317).

results/foundation_models/index.qmd

@@ -0,0 +1,25 @@
+---
+title: "Foundation models"
+subtitle: "Modelling of single-cells to perform multiple tasks."
+image: thumbnail.svg
+page-layout: full
+css: ../_include/task_template.css
+engine: knitr
+fig-cap-location: bottom
+citation-location: document
+toc: false
+---
+
+Recent developments in deep-learning have led to the creation of several "foundation models" for single-cell data.
+These are large models that have been trained on data from millions of cells and am to fully capture the variability in the single-cell landscape.
+Typically, they use a transformer architecture and undergo self-supervised pre-training using masking of parts of the input data.
+Trained foundation models can then be applied to a variety of downstream tasks, either by directly feeding new data into the model or by fine-tuning to better fit a new dataset or to produce a specific output.
+The general nature of single-cell foundation models and the large amount of data they have been trained on makes them potentially powerful tools for single-cell analysis but their performance is yet to be fully established.
+
+Open Problems plans to build on the existing evaluation of foundation models by incorportating them into our continuous benchmarking framework.
+Models will be included in existing tasks where they will be compared to more traditional methods.
+We also plan to compare foundation models across tasks to evaluate their flexibility and compare them to each other.
+
+If you are interested in evaluating foundation models for single-cell data please fill in the form below to get in touch.
+
+<iframe src="https://docs.google.com/forms/d/e/1FAIpQLSeB6ZHTL8yvKpA8VNRoQY28p5WDpi2WEtwRkeGSgRz3Mqw8bA/viewform?embedded=true" width="640" height="959" frameborder="0" marginheight="0" marginwidth="0">Loading…</iframe>