Research project Graphs and graph transformation reveal how ideas, systems, and organisations are connected and evolve over time. At the same time, neural networks – computer programmes that learn patterns from large amounts of data – have become a powerful tool within AI. The drawback of these methods is that they lack the transparency that characterises graph transformation. This project explores how both concepts can be combined to create technology that can both learn and remain comprehensible and reliable.
The aim of the project is to integrate neural methods into rule‑based systems that operate on graphs. The goal is to achieve greater flexibility and learning capability while retaining the transparency and strong algorithmic properties that characterise rule‑based approaches.
Rule‑based systems that perform symbolic computations on graph‑structured data have been studied for more than five decades and are extremely well understood. Due to their high level of abstraction, they provide transparency: both the computations themselves and their results can be interpreted by humans. They also provide algorithmic transparency, due to a rich theory that has been developed over half a century, as well as strong resource efficiency. However, their discrete nature makes them inflexible and difficult to learn from data.
In contrast, modern neural models of computation are untransparent "black boxes" whose creation is resource-demanding, but they come equipped with powerful machine learning algorithms and can flexibly adapt to unanticipated inputs. This project's aim is to integrate the latter into the former to combine their advantages, thus making rule-based methods more general and adaptive without losing transparency and their good algorithmic properties.
The foundation of the project will be my ongoing work on rule‑based formalisms for the generation and transformation of graphs, which will be taken to a new level by integrating neural methods in two ways: (1) by providing formal hooks which can be instantiated by neural models to restrict or generalize the behavior of the rule-based model and (2), by on-the-fly creation of rules similar to the original rules of a given rule-based system.
The project Neuro‑symbolic graph transformation is funded by the Swedish Research Council, which has awarded just over SEK 5 million to the research. The project is led by Frank Drewes, Professor at the Department of Computing Science, Umeå University.
Please contact Frank Drewes using the details below, where you will also find further information about the Department of Computing Science and the research carried out at Umeå University.
