About
Building autonomous artificial agents able to grow open-ended repertoires of skills across their lives is one of the fundamental goals of AI. To that end, a promising developmental approach recommends the design of intrinsically motivated agents that learn new skills by generating and pursuing their own goals–autotelic agents. However, despite recent progress, existing algorithms still show serious limitations in terms of goal diversity, exploration, generalization or skill composition.
Inspired by the seminal work of the developmentalist Vygotsky, we suggest the design of Vygotskian autotelic agents (VAA) able to interact with others and, more importantly, able to internalize these interactions within the agent to transform them into cognitive tools supporting the development of new cognitive functions. Indeed, most of our skills could not be learned in isolation. Formal education teaches us to reason systematically, books teach us history, and YouTube might teach us how to cook. Most importantly, our values, traditions, norms and most of our goals are cultural in essence. This knowledge, and some argue, some of our highest cognitive functions such as abstraction, compositional imagination or relational thinking, are formed through linguistic and cultural interactions with others.
Therefore, Vygotskian autotelic agents should be immersed into rich socio-cultural worlds, an immensely important attribute of our environment that is mostly omitted in modern AI, including deep reinforcement learning research. We focus on language especially, and how its structure and content may support the development of new cognitive functions in artificial agents, just like it does in humans. To know more, check the perspective paper.
On this webpage, we provide a list of papers connected to VAAI. We also provide a list of inter-desciplinary background papers here
Main Contributors
- Tristan Karch, INRIA.
- Cédric Colas, INRIA.
Contributing
This website is maintained by the research community. If you wish to add a reference or change something on the website, please create a pull request implementing these changes. The sources of the website are available here.