ALMA was represented by Algebraic AI at Conexión AIHUB-Industria, an event organized by the CSIC (Spanish National Research Council) for companies and researchers to discuss, promote and consolidate public-private collaboration in R&D&I in artificial intelligence.
This white paper explains, in simple terms, the main ideas behind Algebraic Machine Learning (AML) and provides insights on how and why it works. AML is a symbolic method that is good for reasoning and has the advantage of being able to learn from data. AML can use continuous input and output, can deal with uncertainty and can combine learning from both, data and formulas. These unique properties show that the main limitations of symbolic methods can be overcome and open a path to a more transparent, trustworthy and understandable AI.
Read more: Algebraic Machine Learning: a new program for Symbolic AI
eProsima is pleased to announce the ongoing work on the AML Integrating Platform (AML-IP) under the scope of the ALMA project.
Read more: Making Algebraic Machine Learning node communication easy with AML Integrating Platform
In this post we will briefly outline our understanding of human-centered and Explainable AI (xAI) and the differences and opportunities we see. We will then present our preliminary work on an algebraic machine learning approach as an example to combine both.
The TU Kaiserslautern team is designing an advanced new hardware accelerator for algebraic machine learning.
Read more: Building hardware architectures for Algebraic Machine Learning