Update slides

Author: nicolasreichardt

notebooks/try_few_shot.ipynb

@@ -1129,7 +1129,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.11.14"
+   "version": "3.10.19"
   }
  },
  "nbformat": 4,

presentation/tutorial-new-tutorial-group-1.html

@@ -74,32 +74,68 @@ format:
 
 ## Few Shot Learning in General
 
-tbd
+### Few-Shot Learning (FSL)
+- Learning new **tasks, labels, or segmentations** from very few labeled examples
+  *(N-way, K-shot)*
+- **Motivation**:
+  - Data scarcity
+  - Expensive and time-consuming annotation
 
+### Few-Shot Semantic Segmentation (FSSS)
+- **Goal**: Segment novel object classes using only a few annotated examples
+- Assigning a class label to **every pixel**
+
+
+---
+
+### Prototypical Networks (ProtNets)
+- Learn a shared **embedding space**
+- Pixels belonging to the same class are **close in feature space**
+- Class representations are formed as **prototypes**
+- Training follows an **episodic framework**
+- Each episode consists of:
+  - **Support set**:
+    - Few images with **pixel-level masks**
+    - Defines the target classes
+  - **Query image**:
+    - Image where the model must segment the target classes
 
 
-## Prototypical Network
 
-![](figures/illustration_prototypical_network.png){width=100%}
 
-(modified figure from paper) [SRPNet](https://arxiv.org/abs/2210.16829)
 
-- high-level schematic (support → prototype → similarity → segmentation)
 
-- 1-way-1 shot --> explain what it means
 
-- Data Preprocessing (e.g. Augementation, Geographic Splits)
+## Prototypical Network Overview
 
-- Model Architecture (feature Exctraction, CNN --> Number of Layers, Backbone)
+### Workflow
+- Support Image → Prototype → Similarity → Query Segmentation
 
-- Training strategy
 
-  - Loss function
+### Feature Extraction
+- **Backbone:** ResNet-18 CNN, pretrained on ImageNet
+- **Projection:** feature maps → embedding dimension (256 channels)
 
-  - Evaluation metrics
 
+### Evaluation Metric
+$$
+\mathrm{IoU} = \frac{|A \cap B|}{|A \cup B|}
+$$
 
 
+---
+
+## Prototypical Network Overview
+
+::: {#fig-protnet fig-align="center"}
+![](figures/illustration_prototypical_network.png){width=100%}
+
+*(Modified figure from* [SRPNet](https://arxiv.org/abs/2210.16829)*)*
+:::
+
+
+---
+
 ## (Preliminary) Results
 
 - Show performance for 1-shot / 5-shot / full-data comparison
@@ -110,23 +146,27 @@ tbd
 
 ## Wrap-Up/ Discussion
 
-[GitHub Repo](https://github.com/hertie-data-science-lab/tutorial-new-tutorial-group-1/tree/main)
+**What we still want to work on**:
+
+- Testing different kind of pretrained models as our encoder (ResNet-50, ResNet pretrained on satellite images)
 
+- Play around with different distance metrics (Cosine Similarity and Fidelity)
 
-**What we still need to finalize**:
+- Evaluate different K values and see how they perform
 
-- insert bullet point here
 
-- insert bullet point here
+**Discussion and Key Takeaways**
 
+- Strong 1-shot performance: even with minimal labeled data, results were impressive
 
-**Questions to discuss in class/ lynn**
+- Dataset brought its own challenges (designed primarily for PV assessment, not general segmentation)
 
-- strengths
+- Limited scope: focused only on roofs in Geneva → raises questions of generalizability
 
-- weaknesses
+- More diverse data or complex models could improve performance
 
-- failure cases (shadows, tiny rooftops)
+
+[GitHub Repo](https://github.com/hertie-data-science-lab/tutorial-new-tutorial-group-1/tree/main)