Machine learning studies computer algorithms that improve (learn) automatically by experience. Algebraic machine learning (AML) produces generalizing models from semantic embeddings of data into discrete algebraic structures. It has the potential to flexibly combine unstructured and complex information contained in training data, and has higher mathematical transparency than deep networks and other optimization-based statistical methods.

The ALMA project proposes to innovate this method by adjusting it for interactive possibilities, which can improve its practical use. During the research, which starts in September 2020 and will run for 48 months, until August 2024, the project team wants e.g. to create the groundwork for developing algorithms, methods and systems to better understand the principles of generalization in AML and the ability to combine it with other machine learning techniques.

Another objective is to develop a high-level language for AML, which even non-expert users would use to interact with the AI without the need to master the underlying low-level algebraic descriptions or representations that AML uses.

“The potential benefit to society of these new capabilities is considerable. In key areas such as healthcare, an enormous amount of information is already formalized and available but cannot be used by current machine learning methods. Using formal knowledge (knowledge that is already understood by humans and available in a formal language) can extend the domain of applicability of machine learning to the most important problems of society, including healthcare, decision-making or machine automation,” stated the researchers in the project proposal.

ALMA was chosen in 2019 EIC Pathfinder call (FET Proactive: emerging paradigms and communities, FETPROACT-EIC-05-2019), in the subtopic category of human centric AI. The work on the project has started in September 2020 and will run until August 2024. Among the participants, there are research institutions from Spain, Portugal, Germany, France and Finland.

Background information

FET-Open and FET Proactive are now part of the Enhanced European Innovation Council (EIC) Pilot (specifically the Pathfinder), the new home for deep-tech research and innovation in Horizon 2020, the EU funding programme for research and innovation.