tiny AI: Visual Symbol Grounding with Distributed Representations

Increasing the degree of autonomy of both software and robotic systems through intelligent
behaviours is a great not fully solved technological challenge of the present days. Solving it is
essential for more rapid emergence of intelligent technologies such as self-driving cars,
autonomous robots, large scale information retrieval systems, which will empower strategies for
sustainable development of cities, building resilient infrastructures and smart industries. In many
cases, intelligent behaviour in the AI sense is about integration of methods for acquiring and
processing of visual information and the symbolic methods for representing conceptual
knowledge and inferring from them, i.e. reasoning, modelling, and planning. The stumbling block
for such integration is a yet unsolved fundamental AI problem known as the symbol grounding
problem. This project is dedicated to the development of new methods and algorithms that will
advance solving this fundamental scientific challenge in the context of novel brain-inspired
hardware also known as. neuromorphic computing.

Funder 1: SSF. Budget: 2 000 000 SEK.
Funder 2: Scholars at Risk Sweden via Swedish Research Council. 1 000 000.