Marie-Sophie Pichler
Visual Analytics für Deep Learning mit Graphen: Case Study Neuronen Clustering
[thesis]

Information

  • Publication Type: Master Thesis
  • Workgroup(s)/Project(s): not specified
  • Date: 2024
  • TU Wien Library:
  • Second Supervisor:
  • Open Access: yes
  • First Supervisor: Eduard GröllerORCID iD
  • Pages: 134
  • Keywords: Graph embedding, Visual Analytics

Abstract

Many deep learning applications are based on graph data in order to explore relationships or to analyze structures. Labeling this data is expensive and often requires expert knowledge. For the application of graph clustering to neuron data, the SOTA method GraphDINO generates self-supervised graph embeddings combined with the downstream task of clustering these embeddings. We observe on a particularly challenging neuron dataset that this method does not lead to satisfying clustering results. Therefore we use the graph embeddings generated by GraphDINO as an initial starting point to improve the network and to guide the network training. To achieve this, we developed the visual analytics framework NetDive. The user can analyze the graph embeddings and label single neurons that are falsely clustered. This annotation information is then used to train a semi-supervised model. To this end, we developed a network architecture, named GraphPAWS, that assembles components of GraphDINO and of the semi-supervised network architecture PAWS. The model training can be started from within the visual analytics application NetDive and the resulting graph embeddings are available in NetDive as soon as the retraining is completed. We demonstrate how we iteratively train the model using NetDive and GraphPAWS and evaluate our model against the self-supervised SOTA for our dataset.

Additional Files and Images

Additional images and videos

teaser: Example neurons for each spiny clusters of the BBP dataset, with apical dendrites in lighter color and basal dendrites in darker color. teaser: Example neurons for each spiny clusters of the BBP dataset, with apical dendrites in lighter color and basal dendrites in darker color.

Additional files

thesis
thesis

Weblinks

BibTeX

@mastersthesis{pichler-2024-vaf,
  title =      "Visual Analytics f\"{u}r Deep Learning mit Graphen: Case
               Study Neuronen Clustering",
  author =     "Marie-Sophie Pichler",
  year =       "2024",
  abstract =   "Many deep learning applications are based on graph data in
               order to explore relationships or to analyze structures.
               Labeling this data is expensive and often requires expert
               knowledge. For the application of graph clustering to neuron
               data, the SOTA method GraphDINO generates self-supervised
               graph embeddings combined with the downstream task of
               clustering these embeddings. We observe on a particularly
               challenging neuron dataset that this method does not lead to
               satisfying clustering results. Therefore we use the graph
               embeddings generated by GraphDINO as an initial starting
               point to improve the network and to guide the network
               training. To achieve this, we developed the visual analytics
               framework NetDive. The user can analyze the graph embeddings
               and label single neurons that are falsely clustered. This
               annotation information is then used to train a
               semi-supervised model. To this end, we developed a network
               architecture, named GraphPAWS, that assembles components of
               GraphDINO and of the semi-supervised network architecture
               PAWS. The model training can be started from within the
               visual analytics application NetDive and the resulting graph
               embeddings are available in NetDive as soon as the
               retraining is completed. We demonstrate how we iteratively
               train the model using NetDive and GraphPAWS and evaluate our
               model against the self-supervised SOTA for our dataset.",
  pages =      "134",
  address =    "Favoritenstrasse 9-11/E193-02, A-1040 Vienna, Austria",
  school =     "Research Unit of Computer Graphics, Institute of Visual
               Computing and Human-Centered Technology, Faculty of
               Informatics, TU Wien",
  keywords =   "Graph embedding, Visual Analytics",
  URL =        "https://www.cg.tuwien.ac.at/research/publications/2024/pichler-2024-vaf/",
}