A new AI-assisted algorithm
The last decade has seen an exponential growth in the Artificial Intelligence field of study. The literature is replete with numerous innovations in terms of Knowledge Discovery (KD) going from Computer Vision (Hinton et al, 2007) to Natural Language Processing (Collobert et al, 2008). The last wave in AI research and notably in Deep Networks has reached a level of efficiency that has never been experienced before. AI-based algorithms are now challenging Human Intelligence in many domains. The Go game invented 2500 years ago in China remained one of the few games where Human players were undefeated until Google’s AlphaGo beats Humanity in 2016.
AI research is a cross-cutting area gathering many fields of studies with numerous concrete applications (see Figure 2). It takes on its full meaning for decision problems in which the search space is inaccessible for entire and explicit enumeration. This is notably the case for Image Classification where the number of features/dimensions characterizing an image is beyond the human perception limited to our 3 conventional dimensions.
For this purpose, AI-based algorithms rely on a strong mathematical theory, i.e., the Optimisation Theory which is its main kernel. Human Intelligence heavily relies on the power of human beings to optimise unconsciously when executing tasks. This is the reason why We (Humans) developed machines or algorithms that mimic Us. The potential of an AI-based algorithm strongly depends on its capacity to optimise one or several criteria.
In this project, we propose an innovative approach to generate sustainable and trustworthy recommitment strategies with the aid of AI-based algorithms. Our main attempt is not only to develop an algorithm replacing human strategies but also to design a Sustainable and Trustworthy AI Recommitment System (STAIRS) guiding dynamically the search of recommitment strategies in order to build portfolios of responsible investments. STAIRS will be a cooperative AI-assisted algorithm taking into account human decision makers and their attempts. STAIRS’s main prerogatives will be space exploration of strategies to discover the most suitable given a certain scenario.