GMLR 2022 Track

The 37th ACM/SIGAPP Symposium on Applied Computing (SAC 2022) has been a primary gathering forum for applied computer scientists, computer engineers, software engineers, and application developers from around the world. SAC 2022 is sponsored by the ACM Special Interest Group on Applied Computing (SIGAPP), and will be held in Brno, Czech Republic.

The technical track on Graph Models for Learning and Recognition is at its first edition and will be organized within SAC 2022.

Description
Graphs have gained a lot of attention in the pattern recognition community thanks to their ability to encode both topological and semantic information. Despite their invaluable descriptive power, their arbitrarily complex structured nature poses serious challenges when they are involved in learning systems.

The complexity of graph data carries significant challenges for existing algorithms. Some (but not all) of challenging concerns are: a non-unique representation of data, heterogeneous attributes (symbolic, numeric, etc.), and so on. Furthermore, in the machine learning context, even other important issues are addressed. For example, as graphs can be irregular, a graph may have a variable size of unordered nodes, and the nodes may have a different number of neighbors, resulting in some important operations (e.g., convolutions) that are easy to compute in a vector domain, but difficult to apply to the graph domain.

In recent years, due to their widespread applications, graph-based learning algorithms have gained much research interest. Encouraged by the success of CNNs, a wide variety of methods have redefined the notion of convolution for graphs and provided, in particular, a suitable representation for ubiquitous spatio-temporal data. These new approaches have in general enabled effective training and achieved in many cases better performances than competitors, though at the detriment of computational costs.

Typical examples of applications dealing with graph-based representation are: scene graph generation, point clouds classification, and action recognition in computer vision; text classification, inter-relations of documents or words to infer document labels in natural language processing; forecasting traffic speed, volume or the density of roads in traffic networks, whereas in chemistry researchers apply graph-based algorithms to study the graph structure of molecules/compounds.

This track as a whole intends to focus on all aspects of graph-based representations and models for learning and recognition tasks.

Topics include but are not limited to
  - Graph Neural Networks: theory and applications
  - Deep learning on graphs
  - Graph or knowledge representational learning
  - Graphs in pattern recognition
  - Graph databases and linked data in AI
  - Benchmarks for GNN
  - Dynamic, spatial and temporal graphs
  - Graph methods in computer vision
  - Human behavior and scene understanding
  - Social networks analysis
  - Data fusion methods in GNN
  - Efficient and parallel computation for graph learning algorithms
  - Reasoning over knowledge-graphs
  - Interactivity, explainability and trust in graph-based learning
  - Probabilistic graphical models
  - Biomedical data analytics on graphs