Welcome to the official website of the IEEE Task Force on Learning for Structured Data (LEARN4SD) promoted by the IEEE Technical Committee on Neural Networks.
Structured data (e.g. sequences, trees and graphs) are a natural representation for compound information made of atomic information pieces, i.e. the nodes (and their labels), and their relationships represented by the edges (and their labels).
Graphs are the most general and complex form of structured data, which allow to represent networks of interacting elements, e.g. social or metabolomics graphs, as well as data where topological variations define important and interesting properties, e.g. molecular compounds.
Being able to process data by accounting for these rich structured forms provides a fundamental advantage when it comes to identifying data patterns useful for predictive and/or explorative analyses. This has motivated a recent increasing interest of the machine learning community into the development of learning models for structured information.
Task Force Goals
The broad aim of the LEARN4SD task force is to promote events, research and dissemination activities for the community working on machine learning for structured data. Please regularly check our Activities page.
The detailed goals of the task force are
Create a visible, active and internationally recognized community of top researchers on machine learning for structured data
Contribute to assembling and diffusing reliable and challenging benchmarks to assess and validate learning models on structured data downstream tasks
Foster scientific events and educational-tutorial activities on this topic
Pursue industry engagement, seeking and promoting impacting applications of learning models for structured data
Promote cross-disciplinary collaborations between deep learning researchers and scientists from other fields, such as social science, biology, chemistry, etc..
The LEARN4SD task force focuses on the vertical theme of learning models and downstream tasks on structured data, considering many different paradigms of learning-based approaches including, but not limited, to deep learning and neural networks.
A (non-exhaustive) list of topics of interest follows:
Deep learning and representation learning for graphs and network data
Graph generative models
Adaptive processing of structured data with neural, probabilistic and kernel-based methods
Recurrent, recursive and contextual models
Tensor methods for structured data
Applications including but not limited to: natural language processing, machine vision (e.g. point clouds as graphs), materials science, chemoinformatics, computational biology, social networks.
Please feel free to contact us to discuss and propose further topics!