Image Learning at the Crossroads Between Human and Artificial Intelligence

Techniques for thought representation, facilitating brainstorming and thought-sharing – as for instance Venn-diagrams, flow-charts and mind-mapping – have been used since well before the current Digital Age. Now, image-based digital tools are being perfected to an unprecedented level, providing both humans and machines with a structured and synthetic way to analyse, simulate and design reality. They can even help humans and machines cooperate and learn from each other with enhanced transparency and mutual understanding, often within semantically structured processes. It is for instance the case of Knowledge-Graphs, on which some Artificial Intelligence (AI) algorithms can be based upon. Indeed, old techniques, new technologies and a series of innovative uses thereof are converging so that shared multi-dimensional images and maps of ideas may provide both humans and AI a faster way to grasp, assimilate and cognitively structure reality, and a very effective way to communicate and share thoughts. Moreover, AI is becoming an ideal playground to deepen in parallel – through analogies but also through differences – our knowledge about our own intelligence and learning processes; vice versa, a deeper comprehension of human intelligence can help us better structure AI. New epistemological paradigms may emerge from hybrid processes, and a deeper understanding of intelligence as such may stem therefrom. This paper describes such interplay of visual tools and thoughts, especially in the relationship between and among humans and machines, where images could be considered as the core of a basic common language of cognition, and for new hybrid humans/machines systems to learn about the world.