Summary

• Page properties (3 modified, 0 added, 0 removed)
• Attachments (0 modified, 1 added, 0 removed)
  • image-20240131151228-1.png

Details

Page properties

Title

...	...	@@ -1,1 +1,1 @@
1		-A Machine Learning Approach for Node Size Approximation in Top-down Layout
	1	+Smart Zoom for Edge and Port Labels

Parent

...	...	@@ -1,1 +1,1 @@
1		-For Students.Topics for Student Theses.WebHome
	1	+Theses.Topics for Student Theses.WebHome

Content

...	...	@@ -1,45 +1,35 @@
1		-Top-down Layout is a technique to draw large hierarchical diagrams from the root node downwards, scaling children down to fit in the space provided by their parents. This is in contrast to bottom-up layout where children are laid out first and the parents' dimensions are determined accordingly afterwards.
	1	+One of the challenges of drawing large diagrams is the question of how to maintain a readable diagram. While details remain readable when zoomed in, it is difficult to obtain a good overview while zoomed out, because labels are generally too small to read. In our current tooling in [[klighd-vscode>>https://github.com/kieler/klighd-vscode]] there is a "Smart Zoom" feature for nodes. It hides details within nodes and draws titles of nodes at an increased scale to keep them readable while zooming out. This can be seen in the figure below.
2	2
3		-In top-down layout a strategy needs to be used to set node sizes without knowledge of the hierarchical contents of the node as that has not been processed/laid out at that point. Current strategies are:
	3	+The task for this thesis is to implement a similar feature for edge labels and for port labels. This requires identifying and designing a sensible method for determining the available label bounds in which the label may be upscaled. The work should build upon the existing smart zoom feature.
4	4
5		-* Using a default base size
6		-* Counting the number of children and taking the square root as a multiplication factor for the default base size
7		-* Computing the layout of only the children (look-ahead layout)
	5	+[[image:image-20240131151228-1.png||height="461" width="1007"]]
8	8
9		-The main challenge is to get an approximation that gives a suitable aspect ratio (close to what will actually be required).
10		-
11		-Graphs are complex feature vectors and the solution space is very large without necessarily one correct and optimal answer. Therefore, a machine learning (ML)-based approach may help find good solutions.
12		-
13		-== Example Top-down Layout of an SCChart ==
14		-
15		-[[image:attach:Controller_topdown_v3.png]]
16		-
17	17	= Goals =
18	18
19		-* Use the KiCoDia benchmarking tool to extract feature vectors from existing models
20		-* Train and evaluate an ML model on the extracted data sets
21		-* Integrate the model as a new node size approximator into top-down layout
	9	+* Implement a smart zoom feature for edge labels
	10	+* Implement a smart zoom feature for port labels
	11	+* Evaluation of sensible configurations and usability of features
22	22
23	23	= Scope =
24	24
25		-Master's (Bachelor's) Thesis
	15	+Master's/Bachelor's Thesis
26	26
27	27	= Related Work/Literature =
28	28
29		-[Under Review] M. Kasperowski and R. von Hanxleden, //Top-down Layout: Effectively Utilizing Zoom for Drawings of Compound Graphs//
	19	+* Prior work on existing Smart Zoom feature:
	20	+** [[Master Project>>doc:Projects.Google Maps for Models - Summer Term 2021.WebHome]]
	21	+** Master's Thesis: Bennet Bleßmann, //Google Maps for Models//, April 2022
	22	+* (((
	23	+De Carlo, G., Langer, P., & Bork, D. //Advanced visualization and interaction in GLSP-based web modeling: realizing semantic zoom and off-screen elements//. Proceedings of the 25th International Conference on Model Driven Engineering Languages and Systems (pp. 221-231), October 2022
	24	+)))
30	30
31		-M. Nielsen, //Neural Networks and Deep Learning//, Determination Press, 2015 ([[http:~~/~~/neuralnetworksanddeeplearning.com/index.html>>url:http://neuralnetworksanddeeplearning.com/index.html||shape="rect"]])
32		-
33		-I. Goodfellow and Y. Bengio and A. Courville, //Deep Learning//, MIT Press, 2016 ([[https:~~/~~/www.deeplearningbook.org/>>url:https://www.deeplearningbook.org/||shape="rect"]])
34		-
35	35	= Involved Languages/Technologies =
36	36
37		-* Java / Xtend, Python
38		-* KiCo
39		-* ML Frameworks (to be chosen)
	28	+* Java / Xtend in SCCharts Synthesis
	29	+* Typescript in klighd-vscode
40	40
41	41	= Supervised by =
42	42
43		-Maximilian Kasperowski in cooperation with the [[Intelligent Systems>>url:https://www.ins.informatik.uni-kiel.de/en||shape="rect"]] group.
	33	+Maximilian Kasperowski
44	44
45	45	mka@informatik.uni-kiel.de

image-20240131151228-1.png

Author

...	...	@@ -1,0 +1,1 @@
	1	+XWiki.mka

Size

...	...	@@ -1,0 +1,1 @@
	1	+380.7 KB

Content