docs: sync repository with camera-ready paper

Author: yousefkotp

CHANGELOG.md

@@ -9,6 +9,11 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ### Added
 
+
+## [0.1.0] - 2026-03-31
+
+### Added
+
 - Two-stage training pipeline (self-supervised slide encoding + supervised case-level alignment)
 - CLI interface (`moozy train stage1`, `moozy train stage2`, `moozy encode`)
 - HuggingFace Hub integration for checkpoint downloads

CITATION.cff

@@ -16,6 +16,8 @@ authors:
 repository-code: "https://github.com/AtlasAnalyticsLab/MOOZY"
 url: "https://github.com/AtlasAnalyticsLab/MOOZY"
 license: CC-BY-NC-SA-4.0
+version: 0.1.0
+date-released: 2026-03-31
 keywords:
   - computational pathology
   - digital pathology
@@ -28,7 +30,7 @@ keywords:
   - case-level embedding
   - survival prediction
 preferred-citation:
-  type: article
+  type: conference-paper
   title: "MOOZY: A Patient-First Foundation Model for Computational Pathology"
   authors:
     - family-names: Kotp
@@ -42,4 +44,5 @@ preferred-citation:
       given-names: Mahdi S.
       orcid: "https://orcid.org/0000-0002-9147-0731"
   url: "https://arxiv.org/abs/2603.27048"
+  collection-title: "European Conference on Computer Vision (ECCV)"
   year: 2026

MODEL_CARD.md

@@ -2,6 +2,8 @@
 
 **Paper:** [arXiv:2603.27048](https://arxiv.org/abs/2603.27048)
 
+**Venue:** Accepted to ECCV 2026
+
 ## Model Description
 
 MOOZY is a patient-first foundation model for computational pathology. It produces 768-dimensional embeddings at the slide and case (patient) level from whole-slide image feature grids.
@@ -50,7 +52,9 @@ Anatomical sites: adrenal gland, bladder, brain, breast, cervix, colon/rectum, e
 ## Limitations
 
 - Trained exclusively on public data, which skews toward TCGA/CPTAC demographics (predominantly North American). Performance on underrepresented populations is unknown.
-- Treatment Response (MBC) showed high variance across folds, indicating limited generalization for that specific task.
+- Produces global slide- and case-level embeddings; it does not support dense localization or segmentation outputs.
+- The benefit of explicit cross-slide aggregation has not been isolated on tasks designed to require deeper multi-slide reasoning like report generation.
+- The held-out tasks mix familiar and new cohort distributions, so the reported results are not a pure OOD evaluation of unseen patient populations.
 
 ## Ethical Considerations
 

README.md

@@ -66,9 +66,10 @@ dev = [
 moozy = "moozy.cli.main:main"
 
 [project.urls]
-Homepage = "https://github.com/AtlasAnalyticsLab/MOOZY"
+Homepage = "https://atlasanalyticslab.github.io/MOOZY/"
 Repository = "https://github.com/AtlasAnalyticsLab/MOOZY"
 Documentation = "https://github.com/AtlasAnalyticsLab/MOOZY/tree/main/docs"
+"Paper" = "https://arxiv.org/abs/2603.27048"
 "Model Weights" = "https://huggingface.co/AtlasAnalyticsLab/MOOZY"
 "Bug Tracker" = "https://github.com/AtlasAnalyticsLab/MOOZY/issues"