School for AI

Besides having hard-coded skills, we expect the AI to be able to learn. We will teach the AI new skills in a gradual and guided way in the School for AI which we are now developing.

In School for AI, we first design an optimized set of learning tasks or a “curriculum.” The curriculum teaches the AI useful skills and abilities, so it doesn’t have to discover them on its own. Next, we subject the AI to training. We use the performance of the AI on the learning tasks of the curriculum to improve both the curriculum and hard-coded AI skills.

Curriculum requirements

A good curriculum:

  • Minimizes the time needed for getting the AI into a target state. When the AI is in the target state, it can learn and evolve on its own;
  • Minimizes the effort required for its own creation;
  • Minimizes the number of skills that need to be hard-coded into the AI.

Finding the optimal curriculum for the AI is a multi-objective optimization problem. The better the curriculum, the faster the learning. However, it isn’t possible to design a universally optimal curriculum (recall the no free lunch theorem). We are limited by the level of our current knowledge and by the eventual architecture of the general AI.

However, we believe that a high-quality curriculum can optimize the learning process and allow for rapid advances in AI breakthroughs over purely algorithmic advances.

Other methods of machine learning (e.g. reinforcement learning) can be combined with curriculum learning for improved performance.

Artificial learning environment

For teaching the AI, we have created a simulated visual toy world with simplified physical laws. We are designing our curriculum to teach the AI from the most basic rules of the world to the most complex ones, up to the point where it can start learning on its own.

The goal is not to teach the AI any arbitrary and specific facts about the world, but the opposite: to teach it useful and general skills for a more efficient understanding and exploration of the world, and for better and more general problem-solving.

During the development of the School for AI, we encountered an interesting problem – how should we specify the tasks for the AI? When there is no or very little common language, it is very challenging and time-consuming to explain tasks to the AI. For this reason, we are focusing on early language acquisition. To cut down on AI development time, we want to be able to efficiently communicate with the AI as soon as possible.

Join GoodAI

Are you keen on making a meaningful impact? Interested in joining the GoodAI team?

View open positions